INTERNATIONAL BUSINESS MACHINES CORPORATION patent applications published on December 28th, 2023

Patent applications for INTERNATIONAL BUSINESS MACHINES CORPORATION on December 28th, 2023

VOLTAGE-SWING METHOD FOR CARBON CAPTURE USING POROUS CARBONS (17808572)

Main Inventor

Binquan Luan

Brief explanation

The abstract describes a method and system for capturing carbon dioxide using a voltage-swing approach. Here is a simplified explanation of the abstract:

• The invention proposes a method and system for capturing carbon dioxide from a gas mixture.
• The process involves using a sorbent material that can selectively adsorb carbon dioxide through physisorption.
• To increase the sorbent's selectivity and adsorption capacity, a positive electrical charge is applied to the sorbent.
• This positive charge enhances the sorbent's ability to capture carbon dioxide from the gas mixture.
• Once the carbon dioxide is captured, it can be liberated from the sorbent by removing the positive electrical charge and applying a desorption method.
• The desorption method helps release the carbon dioxide from the sorbent, allowing it to be collected and stored.

Potential applications of this technology:

• Carbon capture and storage systems in power plants and industrial facilities.
• Reduction of greenhouse gas emissions from fossil fuel combustion processes.
• Carbon dioxide capture in natural gas processing and biogas purification.
• Carbon capture in the production of hydrogen and other industrial processes.

Problems solved by this technology:

• The method provides an efficient and selective approach for capturing carbon dioxide from gas mixtures.
• It addresses the need for effective carbon capture technologies to mitigate climate change and reduce greenhouse gas emissions.
• The voltage-swing approach improves the sorbent's selectivity and adsorption capacity, enhancing the overall carbon capture process.

Benefits of this technology:

• Increased efficiency in capturing carbon dioxide from gas mixtures.
• Selective adsorption of carbon dioxide, reducing the capture of other gases.
• Potential for large-scale implementation in various industries.
• Contribution to mitigating climate change by reducing greenhouse gas emissions.

Abstract

A method and system for carbon capture through a voltage-swing is provided. The present invention may include capturing carbon dioxide from a gas mixture through physisorption by applying a positive electrical charge to a sorbent to increase the sorbent's selectivity and adsorption and liberating the carbon dioxide from the sorbent by removing the positive electrical charge from the sorbent and applying a desorption method to the sorbent.

RECYCLING PROCESS FOR THE RECOVERY OF COTTON FROM POLYESTER-COTTON FABRICS AND/OR FIBERS (18244744)

Main Inventor

Gregory Breyta

Brief explanation

The patent application describes a process to obtain polyester-free cotton from fabric and/or fibers containing polyester and cotton. The process involves reacting the fabric and/or fibers with an amine organocatalyst and/or carboxylic acid salt and an alcohol solvent. The reaction results in the recovery of polyester-free cotton, reusable catalyst/salt, polyester monomer, and unreacted alcohol. The process can be conducted in batches or as a continuous flow process and can be applied to various polyester-cotton fabrics and fibers, including those with additional materials like polyethers, polyolefins, polyurethanes, nylon, rayon, acetate, viscose, modal, acrylic, wool, and combinations thereof.

• The process converts polyester-containing fabric and fibers into polyester-free cotton.
• It involves reacting the fabric and/or fibers with an amine organocatalyst and/or carboxylic acid salt and an alcohol solvent.
• The reaction can be conducted in batches or as a continuous flow process.
• The process recovers polyester-free cotton, reusable catalyst/salt, polyester monomer, and unreacted alcohol.
• It can be applied to various polyester-cotton fabrics and fibers, including those with additional materials.

Potential Applications

• Textile industry
• Recycling industry
• Sustainable fashion

Problems Solved

• Eliminates polyester from cotton fabric and fibers
• Provides a method for recovering valuable materials from polyester-cotton blends
• Offers a solution for recycling polyester-containing textiles

Benefits

• Production of polyester-free cotton
• Reusable catalyst/salt
• Recovery of polyester monomer and unreacted alcohol
• Versatile application to various polyester-cotton fabrics and fibers

Abstract

Polyester-free cotton is obtained from a fabric and/or fibers containing polyester and cotton by reacting the fabric and/or fibers with an amine organocatalyst and/or carboxylic acid salt of same and an alcohol solvent. The reaction, which may be run in batches or as a continuous flow process, recovers (i) polyester-free cotton as a solid inert by-product of the reaction, (ii) the amine organocatalyst and/or carboxylic acid salt of same for reuse, (iii) a polyester monomer product, and (iv) unreacted alcohol. The reaction works on any polyester-cotton fabric and/or fibers, including those that have at least one additional material, such as polyethers polyolefins, polyurethanes, nylon, rayon, acetate, viscose, modal, acrylic, wool, and combinations thereof.

SELF HEATING EXTRACTION DESIGN AND METHODOLOGY WITH POWER CONSUMPTION CORRECTION (17848616)

Main Inventor

HUIMEI ZHOU

Brief explanation

The patent application describes a semiconductor device structure and method for accurately determining power consumption and extracting real temperatures due to self-heating effects. This is achieved by using a first transistor device as a heater device and measuring the electrical characteristics of an adjacent second transistor device or junction device.

• The first transistor device is operated at different states and power levels, and the electrical characteristics of the adjacent device are measured at each state.
• The measured electrical characteristics are correlated to the power consumption of the first device, excluding power consumption due to resistance of connected metal layers, vias, or contacts.
• This allows for accurate determination of thermal conductivity, which is important for better Design Technology Co-Optimization (DTCO) and evaluation of device/circuit reliability.

Potential Applications

• Semiconductor device manufacturing
• Integrated circuit design and optimization
• Thermal management in electronic devices

Problems Solved

• Inaccurate determination of power consumption and temperature due to self-heating effects in semiconductor devices
• Difficulty in accurately evaluating thermal conductivity for design and reliability purposes

Benefits

• Accurate determination of power consumption and temperature in semiconductor devices
• Improved Design Technology Co-Optimization (DTCO) for better circuit design and optimization
• Enhanced evaluation of device and circuit reliability

Abstract

A semiconductor device structure and methodology for determining a power consumption of the device structure and to extract accurate real temperatures due to self-heating effects. The semiconductor device structure includes a first transistor device formed as a heater device and an adjacent device such as a second transistor device or a semiconductor junction device. In the method, the first transistor device is operable at different operating states (e.g., off state or at different applied power levels), and at each state, an electrical characteristic is measured at the adjacent second transistor device or junction device. The electrical characteristic measured at the adjacent second semiconductor device is correlated to a power consumption of the first device while excluding a power consumption due to voltage drops due to resistance of connected metal layers, vias or contacts. Accurate thermal conductivity is thus achievable for better Design Technology Co-Optimization (DTCO) and device/circuit reliability evaluation.

STEP BY STEP SENSORY RECIPE APPLICATION (17808113)

Main Inventor

Juel Daniel Raju

Brief explanation

The abstract describes a patent application for a method, computer system, and computer program that assist users in preparing and completing a dish within a social sensory recipe program. The invention involves capturing the aroma of a dish using sensors, identifying aroma profiles based on the captured characteristics, comparing the aroma profiles to a recipe profile, and providing feedback to the user based on the comparison.

• The invention aids users in preparing and completing dishes within a social sensory recipe program.
• It captures the aroma of a dish using sensors, including an olfactory sensor.
• The captured characteristics are used to identify aroma profiles of the dish at different steps of the recipe.
• The identified aroma profiles are compared to the aroma profiles of the recipe profile.
• Feedback on the dish is provided to the user based on the comparison.

Potential Applications

• Enhancing the cooking experience by providing real-time feedback on the aroma of a dish.
• Assisting users in achieving desired aroma profiles for specific recipes.
• Facilitating social interactions and sharing of sensory experiences related to cooking.

Problems Solved

• Difficulty in achieving desired aroma profiles for dishes.
• Lack of real-time feedback on the aroma during the cooking process.
• Limited ability to share and discuss sensory experiences related to cooking.

Benefits

• Improved cooking outcomes by providing feedback on the aroma of a dish.
• Enhanced user experience through real-time monitoring and guidance.
• Increased social engagement and sharing of sensory experiences in cooking communities.

Abstract

According to one embodiment, a method, computer system, and computer program product for aiding a user in the preparation and completion of a dish within a social sensory recipe program are provided. The present invention may include capturing one or more characteristics comprising an aroma of a dish associated with a recipe using one or more sensors, wherein the one or more sensors comprise an olfactory sensor; identifying, at two or more steps of the recipe comprising a recipe profile, one or more aroma profiles of the dish based on the captured one or more characteristics comprising the aroma; comparing the one or more aroma profiles of the dish to one or more aroma profiles comprising the recipe profile; and responsive to the comparing, providing feedback of the dish to a user.

THERMAL MANAGEMENT OF COMPUTER HARDWARE MODULES (17849679)

Main Inventor

Yuanchen Hu

Brief explanation

The patent application describes a structure for managing the heat generated by pluggable hardware modules, such as optical transceivers. The structure includes a socket assembly that can receive the hardware module and has an integrated fluid reservoir containing a thermal interface material (TIM). When the hardware module is plugged into the socket assembly, a gap is created. The structure also includes dispensing ports that automatically distribute the TIM from the reservoir into the gap.

• The structure includes a socket assembly that can receive a hardware module.
• The socket assembly has an integrated fluid reservoir containing a thermal interface material (TIM).
• When the hardware module is plugged into the socket assembly, a gap is created.
• Dispensing ports are included in the structure to automatically distribute the TIM from the reservoir into the gap.

Potential applications of this technology:

• Thermal management of pluggable hardware modules, such as optical transceivers.
• Cooling of electronic components in various devices.

Problems solved by this technology:

• Heat generated by pluggable hardware modules can cause performance issues and damage to the components.
• Traditional cooling methods may not be efficient or effective for pluggable hardware modules.

Benefits of this technology:

• Improved thermal management of pluggable hardware modules.
• Enhanced performance and reliability of electronic components.
• Simplified cooling process for pluggable hardware modules.

Abstract

Aspects of the disclosure include a structure for thermal management of pluggable hardware modules, an optical transceiver, and a method of cooling a pluggable hardware module. One embodiment of the structure may comprise a socket assembly adapted to receive a hardware module. The socket assembly may comprise an integrated fluid reservoir containing a thermal interface material (TIM). The socket assembly may be further adapted to define a gap when the hardware module is plugged into the socket assembly. The structure may further comprise at least one dispensing port in fluid communication with the integrated fluid reservoir and the gap. The at least one dispensing port may be adapted to automatically distribute TIM from the integrated fluid reservoir into the gap when the hardware module is plugged into the socket assembly.

ANALOG PERSISTENT CIRCUIT FOR STORAGE ACCESS MONITORING (17809184)

Main Inventor

Krishna Thangaraj

Brief explanation

The abstract describes a memory system that includes a memory controller, an analog persistent circuit, a command/address bus, and a memory array. The analog persistent circuit is responsible for monitoring access to the memory and storing persistent data related to memory access counts.

• The memory system includes a media controller and an analog persistent circuit.
• The analog persistent circuit is connected to the media controller and monitors access to the memory.
• The analog persistent circuit stores persistent data related to memory access counts.
• The access signals from the command/address bus are used by the analog persistent circuit to monitor memory access.
• The memory array is connected to the command/address bus and the media controller.

Potential Applications:

• This memory system can be used in various electronic devices that require efficient and reliable memory access monitoring, such as computers, smartphones, and IoT devices.
• It can be utilized in data centers and server farms to ensure optimal memory usage and performance.
• The memory system can be integrated into automotive systems to monitor memory access in vehicles.

Problems Solved:

• The memory system solves the problem of efficiently monitoring memory access in a reliable and persistent manner.
• It addresses the need for accurate data on memory access counts to optimize memory usage and performance.
• The system solves the challenge of storing persistent data related to memory access in a memory system.

Benefits:

• The memory system provides accurate and reliable monitoring of memory access, allowing for efficient memory usage and performance optimization.
• It offers persistent storage of memory access data, ensuring that the information is not lost during power cycles or system resets.
• The system is flexible and can be integrated into various electronic devices and systems, providing a versatile solution for memory access monitoring.

Abstract

A memory system for storage access monitoring is provided. The memory system includes a media controller of a memory. An analog persistent circuit is coupled to the media controller and configured to monitor access to the memory. The analog persistent circuit stores persistent data related to memory access counts access signals from the command/address bus. A command/address bus is coupled to the analog persistent circuit. A memory array is communicatively coupled to the command address and the media controller.

COMPUTER-BASED VERIFICATION OF MODULAR CORRECTIONS OF MULTIPLE LINEAR OPERATIONS IN PARALLEL (17808631)

Main Inventor

Yvo Thomas Bernard Mulder

Brief explanation

The abstract describes a method for generating test data to verify a modular correction of a modular multiplication performed by a multiplier unit for very wide operands. The method involves performing a modular multiplication by correcting a binary multiplication of two operands using a coarse-grained and a fine-grained correction. 

• The multiplier unit performs a modular multiplication by correcting a binary multiplication of two operands using a coarse-grained and a fine-grained correction.
• The computer selects adjacent intervals of the intermediate result and defines a sub-interval closely around a boundary between the adjacent intervals.
• The computer selects a value in the sub-interval and uses a factorization algorithm to determine operands A' and B' for the value V.
• The computer repeatedly determines A' plus varying ε-values as A" values and determines B" values to ensure the modular multiplication corrected by the coarse-grained correction is in the sub-interval.

Potential applications of this technology:

• Testing and verification of modular multiplication performed by multiplier units for very wide operands.
• Development and improvement of modular correction algorithms for efficient and accurate modular multiplication.

Problems solved by this technology:

• Ensures the accuracy and correctness of modular multiplication performed by multiplier units for very wide operands.
• Provides a method for generating test data to verify the modular correction of modular multiplication.

Benefits of this technology:

• Increases the reliability and efficiency of modular multiplication in computer systems.
• Enables the development of more advanced and accurate modular correction algorithms.
• Facilitates the testing and verification process for modular multiplication in multiplier units.

Abstract

Generation of test data for verifying a modular correction of a modular multiplication performed by a multiplier unit for very wide operands includes performing, by a multiplier unit using a computer, a modular multiplication by correcting a binary multiplication of two operands by a coarse-grained and a fine-grained correction. The computer selects adjacent intervals of the intermediate result, defines a sub-interval closely around a boundary between the adjacent intervals, and selects a value in the sub-interval. Moreover, the computer uses a first factorization algorithm for the value V for determining operands A′, B′, where the modular multiplication result of the operands corrected by the coarse-grained correction is in the sub-interval. The computer repeatedly determines A′ plus varying ε-values as A″ values, and determines B″ values, so that the modular multiplication corrected by the coarse-grained correction is in the sub-interval.

MODERNIZING APPLICATION COMPONENTS TO REDUCE ENERGY CONSUMPTION (17809061)

Main Inventor

FAN JING Meng

Brief explanation

The patent application describes a method, computer system, and computer program for application modernization. The invention involves analyzing operation data of an application, identifying entities and interactions within the application, and determining an energy consumption pattern based on workload. The modernization scope of the application is then determined based on the energy consumption pattern.

• The invention analyzes operation data of an application.
• It identifies entities and interactions within the application.
• It determines the energy consumption pattern based on workload.
• It determines the modernization scope of the application based on the energy consumption pattern.

Potential Applications

• This technology can be applied to various software applications that require modernization.
• It can be used in industries such as finance, healthcare, and manufacturing to improve the efficiency and performance of legacy applications.

Problems Solved

• The technology solves the problem of identifying entities and interactions within an application, which helps in understanding the application's structure and dependencies.
• It addresses the challenge of determining the energy consumption pattern of an application, which is crucial for optimizing energy usage and reducing costs.
• The technology solves the problem of determining the modernization scope of an application, providing guidance on which parts of the application need to be updated or replaced.

Benefits

• The technology enables organizations to modernize their applications more effectively by providing insights into their structure and energy consumption.
• It helps in optimizing energy usage, leading to cost savings and environmental benefits.
• The modernization scope determination helps in prioritizing and planning the modernization efforts, saving time and resources.

Abstract

A method, computer system, and a computer program product for application modernization is provided. The present invention may include receiving operation data related to an application. The present invention may include identifying a plurality of entities based on the operation data related to the application. The present invention may include identifying one or more interactions between each of the plurality of entities. The present invention may include determining an energy consumption pattern for the application by analyzing a consumption of energy related to workload. The present invention may include determining a modernization scope for the application based on the energy consumption pattern.

NODE MANAGEMENT FOR A CLUSTER (17808864)

Main Inventor

Hai Hui Wang

Brief explanation

The patent application describes a method, device, and computer program for managing a cluster of computing nodes. Here is a simplified explanation of the abstract:

• The invention involves grouping a cluster of computing nodes into a hierarchy of groups based on grouping policies.
• Each group in the hierarchy has a leader node, which is responsible for collecting and reporting the status of all computing nodes in that group.
• The leader node of a lower-level group reports the status to the leader node of the higher-level group in the hierarchy.

Potential Applications:

• This technology can be applied in various distributed computing systems, such as cloud computing platforms or data centers.
• It can be used to efficiently manage and monitor large clusters of computing nodes, ensuring smooth operation and resource allocation.

Problems Solved:

• Managing and monitoring a large number of computing nodes in a cluster can be challenging and time-consuming.
• This technology solves the problem of efficiently collecting and reporting the status of computing nodes in a hierarchical manner, reducing the complexity of node management.

Benefits:

• The hierarchical grouping and reporting approach simplifies the management of large clusters, making it easier to monitor and control the computing nodes.
• It improves the efficiency of node management by reducing the amount of data that needs to be processed and reported.
• The technology enhances the overall stability and performance of the cluster by ensuring timely and accurate status updates.

Abstract

Disclosed are a computer-implemented method, a device and a computer program product of node management for a cluster of a cluster of computing nodes. A plurality of computing nodes in a cluster can be grouped into a hierarchy of groups according to a hierarchy of grouping policies. One of computing nodes in each group of the hierarchy of groups can be determined as a leader node of the corresponding group. A leader node of a first group can be responsible for collecting and reporting status of all computing nodes in the first group to a leader node of a second group superior to the first group by one level in the hierarchy of groups.

DATA CENTER WITH ENERGY-AWARE WORKLOAD PLACEMENT (17809284)

Main Inventor

Asser Nasreldin Tantawi

Brief explanation

The abstract describes a computer-implemented method, system, and program that aims to improve efficiency by placing workloads on computer servers based on their energy profiles and power consumption data. The method involves several steps:

• Obtaining energy profiles and power consumption data for multiple computer servers.
• Determining the optimal temperature for each server based on its energy profile.
• Determining the target processor utilization for each server based on the optimal temperature.
• Calculating an efficiency rank for each server based on the target processor utilization and power consumption data.
• Deploying workloads on the server with the highest efficiency rank.

Potential applications of this technology:

• Data centers: This method can be used to optimize the placement of workloads in data centers, leading to improved energy efficiency and cost savings.
• Cloud computing: By efficiently placing workloads on servers, cloud service providers can enhance their resource utilization and reduce energy consumption.
• High-performance computing: This method can be beneficial in optimizing the allocation of computational tasks in supercomputers, leading to improved performance and energy efficiency.

Problems solved by this technology:

• Inefficient workload placement: Traditional methods may not consider energy profiles and power consumption data when assigning workloads to servers, resulting in suboptimal resource utilization and higher energy consumption.
• Overheating and underutilization: By determining the optimal temperature and target processor utilization for each server, this method helps prevent overheating while maximizing the utilization of computing resources.

Benefits of this technology:

• Improved energy efficiency: By considering energy profiles and power consumption data, workloads can be placed on servers in a way that minimizes energy consumption.
• Cost savings: Optimizing workload placement can lead to reduced energy costs for data centers and cloud service providers.
• Enhanced performance: By allocating workloads based on optimal temperature and target processor utilization, this method can improve the overall performance of computer servers.

Abstract

A computer-implemented method, a computer system and a computer program product boost efficiency through energy-aware workload placement. The method includes obtaining an energy profile for a plurality of computer servers and power consumption data for each computer server in the plurality of computer servers. The method also includes determining an optimal temperature for each computer server in the plurality of computer servers based on the energy profile. The method further includes determining a target processor utilization for each computer server in the plurality of computer servers based on the optimal temperature. In addition, the method includes calculating an efficiency rank for each computer server in the plurality of computer servers based on the target processor utilization and the power consumption data. Lastly, the method includes deploying a workload on a computer server with a highest efficiency rank.

SYSTEM LOG PATTERN ANALYSIS BY IMAGE SIMILARITY RECOGNITION (17808555)

Main Inventor

Chuan Li

Brief explanation

The present invention is a system that uses artificial intelligence to identify and analyze error patterns in log databases. Here is a simplified explanation of the patent application:

• The system identifies a log database with known error patterns.
• It maps these known error patterns into a series of diagrams.
• An artificial intelligence algorithm is trained using these diagrams.
• The algorithm compares diagrams and returns a distance value.
• The system can then identify logs for a new error pattern.
• It maps the new error pattern to a two-dimensional diagram.
• The algorithm compares the new diagram to the existing diagrams.
• If the distance value is below a threshold, log data is identified.

Potential applications of this technology:

• IT support: The system can help IT support teams quickly identify and troubleshoot errors in log databases.
• Software development: Developers can use the system to analyze and fix bugs in their software.
• Network monitoring: The system can be used to detect and analyze errors in network logs.

Problems solved by this technology:

• Manual error pattern identification: The system automates the process of identifying error patterns in log databases, saving time and effort.
• Efficient troubleshooting: By quickly identifying and analyzing error patterns, the system helps in troubleshooting and resolving issues more efficiently.

Benefits of this technology:

• Time-saving: The system automates the process of identifying and analyzing error patterns, saving valuable time for IT support teams and developers.
• Accuracy: The artificial intelligence algorithm provides accurate comparisons and identifies log data based on distance values, reducing the chances of human error.
• Scalability: The system can handle large log databases and efficiently analyze new error patterns as they arise.

Abstract

The present invention may include an embodiment that identifies a log database with known error patterns. The embodiment may map the known error patterns into plurality of diagrams. The embodiment may train an artificial intelligence (AI) algorithm with the plurality of diagrams, where the AI algorithm compares diagrams and returns a distance value. The embodiment may identify logs for a new error pattern. The embodiment may map the new error pattern to a two-dimensional diagram. The embodiment may compare the two-dimensional diagram to the plurality of diagrams using the AI algorithm and identify log data of the at least one diagram based on determining a distance value is below a threshold for at least one diagram within the plurality of diagrams

PULSED STARK TONES FOR COLLISION MITIGATION (18461589)

Main Inventor

Isaac Lauer

Brief explanation

Techniques for using stark tone pulses to mitigate cross-resonance collision in qubits are presented. A tone management component can control application of pulses to qubits by a tone generator component to mitigate undesirable frequency collisions between qubits. The tone generator component (TGC) can apply an off-resonant tone pulse to a qubit during a gate to induce a stark shift. TGC can apply a cross-resonance tone pulse to a control qubit at a frequency associated with the qubit, wherein the frequency can be stark shifted based on the off-resonant tone pulse. The qubit can be a target qubit, the control qubit itself, or a spectator qubit that can be coupled to the target qubit or the control qubit. The gate can be a cross-resonance gate, a two-qubit gate, or a measurement gate that can utilize an echo sequence, a target rotary, or active cancellation.

• Tone management component controls application of pulses to qubits to prevent frequency collisions.
• Tone generator component applies off-resonant tone pulses to induce stark shift in qubits during gates.
• Cross-resonance tone pulses are applied to control qubits at a frequency associated with the qubit.
• The frequency of the control qubit can be stark shifted based on the off-resonant tone pulse.
• Qubit can be a target qubit, control qubit, or spectator qubit coupled to the target or control qubit.
• Gates can be cross-resonance gates, two-qubit gates, or measurement gates utilizing echo sequence, target rotary, or active cancellation.

Potential Applications

• Quantum computing
• Quantum information processing
• Quantum communication

Problems Solved

• Mitigates undesirable frequency collisions between qubits
• Reduces cross-resonance collision in qubits

Benefits

• Improved performance and reliability of qubits
• Enhanced control over qubit operations
• Enables more efficient quantum computing and information processing

Abstract

Techniques for using stark tone pulses to mitigate cross-resonance collision in qubits are presented. A tone management component can control application of pulses to qubits by a tone generator component to mitigate undesirable frequency collisions between qubits. The tone generator component (TGC) can apply an off-resonant tone pulse to a qubit during a gate to induce a stark shift. TGC can apply a cross-resonance tone pulse to a control qubit at a frequency associated with the qubit, wherein the frequency can be stark shifted based on the off-resonant tone pulse. The qubit can be a target qubit, the control qubit itself, or a spectator qubit that can be coupled to the target qubit or the control qubit. The gate can be a cross-resonance gate, a two-qubit gate, or a measurement gate that can utilize an echo sequence, a target rotary, or active cancellation.

Remote Access Array (17808271)

Main Inventor

Ram Sai Manoj Bamdhamravuri

Brief explanation

The abstract describes a computer system and method for implementing a remote access array. The system includes a processor chip, a main memory region, and a non-addressable memory region containing the remote access array. The remote access array tracks data portions pulled from the main memory region and sent to an external node. It is backed up in the main memory region and can be scrubbed when a multi-drawer working partition is shrunk to fit within a single drawer.

• The system includes a processor chip, main memory region, and non-addressable memory region.
• The non-addressable memory region contains a remote access array.
• The remote access array tracks data portions pulled from the main memory region and sent to an external node.
• The remote access array is backed up in the main memory region.
• When a multi-drawer working partition is shrunk to fit within a single drawer, the remote access array can scrub all of its entries.

Potential Applications

• Remote access arrays can be used in various computer systems and networks.
• This technology can be applied in cloud computing environments where remote access to data is crucial.
• It can be used in distributed systems where data needs to be tracked and backed up.

Problems Solved

• The remote access array solves the problem of efficiently tracking and managing data portions pulled from the main memory region.
• It addresses the need for reliable backup and scrubbing of the remote access array.
• The technology solves the challenge of shrinking a multi-drawer working partition without losing important data tracking information.

Benefits

• The system provides efficient tracking and management of data portions sent to external nodes.
• The backup feature ensures data integrity and reliability.
• The ability to scrub all entries when shrinking a working partition helps maintain data consistency and efficiency.

Abstract

A computer system and a method implementing a remote access array are provided. A first drawer may include a first processor chip. A first main memory region may be operatively connected to the first processor chip. A first non-addressable memory region may be operatively connected to the first processor chip and may include the first remote access array. The first remote access array may be configured to track data portions that are pulled from the first main memory region and that are sent to an external node. The first remote access array may be backed up in the first main memory region. The first remote access array may include one or more entries and may be configured to scrub all of the entries in response to a multi-drawer working partition being shrunk to fit within the first drawer.

PREDICTION OF FILE INTERACTION BY A USER (17809296)

Main Inventor

Jing Zhao

Brief explanation

The abstract describes a computer-implemented method, program, and system for automatically performing file management operations. Here is a simplified explanation of the abstract:

• The technology automatically monitors a group of files and generates tracking attributes for each file.
• It tracks user interactions with the files and generates prediction vectors based on these interactions.
• The system then identifies files with tracking attributes that correlate with the prediction vectors.
• Finally, it performs specific operations on the identified files based on the prediction vectors.

Potential Applications

This technology has potential applications in various fields, including:

• Data management systems
• File organization and categorization tools
• Automated file backup and synchronization software
• Content management systems

Problems Solved

The technology addresses several problems related to file management, such as:

• Manual file organization and categorization, which can be time-consuming and prone to errors.
• Difficulty in identifying relevant files based on user interactions.
• Inefficient file backup and synchronization processes.

Benefits

The use of this technology offers several benefits, including:

• Time savings by automating file management tasks.
• Improved organization and categorization of files.
• Enhanced user experience by predicting and performing relevant file operations.
• Efficient file backup and synchronization processes.

Abstract

A computer-implemented method, computer program product and computer system to automatically perform file management operations is provided. A processor identifies a plurality of files to monitor. A processor generates tracking attributes for the plurality of files. A processor monitors user interactions with the plurality of files. A processor generates prediction vectors for a plurality of file interactions based on the user interactions with the plurality of files. A processor determines at least one file in the plurality of files with tracking attributes that correlate with at least one prediction vector. A processor performs an operation on the at least one file that corresponds with the at least one prediction vector.

ADAPTIVE MESSAGE RETENTION TIME BASED ON USER INTERACTION (17809072)

Main Inventor

Klaus Rindtorff

Brief explanation

The present invention is a computer system and method for dynamically adjusting message retention times in a messaging system. It involves receiving and storing messages, determining interactions with the stored messages, and calculating a retention score for each message based on the type and frequency of interactions. Messages with a retention score below a predetermined threshold are then deleted.

• Computer-implemented method for dynamically adjusting message retention times in a messaging system
• Receive and store a plurality of messages in a storage system
• Continuously determine interactions with the stored messages
• Calculate a retention score for selected messages based on weighted interaction type count values
• Delete messages with a retention score below a predetermined threshold

Potential applications of this technology:

• Messaging systems and platforms
• Email servers and clients
• Social media platforms
• Chat applications

Problems solved by this technology:

• Efficient management of message storage
• Reducing clutter and unnecessary data in messaging systems
• Improving system performance and response time

Benefits of this technology:

• Optimize storage space by deleting low-value messages
• Improve user experience by reducing clutter and improving system performance
• Enhance system efficiency by dynamically adjusting retention times based on user interactions

Abstract

Embodiments of the present invention provide computer-implemented methods, computer program products, and computer systems. Specifically, embodiments of the present invention can dynamically adjust individual message retention times in a messaging system is disclosed. Embodiments of the present invention can then receive a plurality of messages, store the plurality of messages in a storage system, determine continuously, by the messaging system, interactions with messages stored by the messaging system, and determine dynamically, by the messaging system, a retention score value for selected messages based on a sum of weighted interactions type count values of the determined interactions. Finally, embodiments of the present invention can delete by the messaging system, messages for which the retention score value is below a predetermined retention threshold value.

QUESTION-ANSWERING SYSTEM FOR AUTONOMIC MANAGEMENT (17846115)

Main Inventor

MUDHAKAR SRIVATSA

Brief explanation

The patent application describes a method, system, and computer program for automatically managing a data processing system (DPS). The method involves receiving operations data associated with the DPS and using a Question-Answer (QA) system to generate a question sentence from the operations data. A window of the operations data is sampled, and key performance indicators (KPI) are extracted from the sampled window and provided to the QA system.

• The method involves receiving operations data from a data processing system.
• A Question-Answer (QA) system is used to generate a question sentence from the operations data.
• A window of the operations data is sampled.
• Key performance indicators (KPI) are extracted from the sampled window.
• The extracted KPIs are provided to the QA system.

Potential Applications

This technology has potential applications in various industries and sectors, including:

• IT and technology companies that need to manage and optimize their data processing systems.
• Service providers that offer data processing services and need automated management tools.
• Companies that rely on data processing systems for their operations and want to improve efficiency and performance.

Problems Solved

The technology addresses the following problems:

• Manual management of data processing systems can be time-consuming and prone to errors.
• Identifying key performance indicators (KPIs) from operations data can be challenging and require manual analysis.
• Lack of automated tools for managing and optimizing data processing systems can lead to inefficiencies and performance issues.

Benefits

The technology offers the following benefits:

• Automation of data processing system management, reducing the need for manual intervention.
• Efficient extraction of key performance indicators (KPIs) from operations data.
• Improved performance and optimization of data processing systems.
• Time and cost savings through automated management and optimization processes.

Abstract

Disclosed are methods method, systems, and computer program products for automatic management of a data processing system (DPS). One embodiment of the method may comprise receiving, at a computer processor, operations data associated with a DPS; and generating, by a Question-Answer (QA) system executing on the computer processor, a Question sentence from the operations data. The method may further comprise sampling a window of the operations data, extracting a plurality of key performance indicators (KPI) from the operations data in the sampled window, and providing the plurality of KPI to the QA system.

UNIFIED DATA CLASSIFICATION TECHNIQUES (17809034)

Main Inventor

Youngja Park

Brief explanation

The patent application describes a method, computer system, and computer program for data processing. The process involves obtaining multiple files from a data source and analyzing them to gather information about their content and structure. 

• The files are analyzed to determine the structural information and content details of each file.
• The information in each file is sorted and categorized based on common content.
• Sensitive information is extracted and categorized separately.
• The categorized information is then merged using the categories to create a single unified file.

Potential applications of this technology:

• Data management and organization in large-scale databases.
• Content analysis and categorization in document management systems.
• Information extraction and classification in data mining and analysis tasks.

Problems solved by this technology:

• Efficiently organizing and categorizing large volumes of data.
• Identifying and extracting sensitive information from files.
• Streamlining the process of merging and unifying data from multiple sources.

Benefits of this technology:

• Improved data organization and accessibility.
• Enhanced efficiency in analyzing and processing large amounts of data.
• Enhanced security and privacy by identifying and categorizing sensitive information.

Abstract

A method, computer system, and a computer program product for data processing, comprising obtaining a plurality of files from a data source. These files are analyzed the files for information about the content and in order to determine structural information of each file. Once the files have been analyzed, information in each file may be sorted and categorized by common content. Sensitive information may also be extracted and categorized separately. Information may then be then merged using the categories to create a single unified file.

Dynamic Threshold-Based Records Linking (17808740)

Main Inventor

Abhishek Seth

Brief explanation

The patent application describes a method for linking records in a database. Here are the key points:

• The method selects two records from a customer's records for comparison.
• It determines if the two records belong to different entities.
• If they do, it identifies a local auto-link-threshold value for the different entities.
• The method then performs an attribute comparison between the two records.
• Based on the attribute comparison, it generates a comparison score.
• If the comparison score is higher than the local auto-link-threshold value, the two records are linked.

Potential applications of this technology:

• Customer relationship management systems
• Data integration and consolidation platforms
• Fraud detection systems
• Identity verification systems

Problems solved by this technology:

• Efficiently linking records from different entities in a database
• Reducing manual effort in record linking and data integration processes
• Improving accuracy and reliability of record linking

Benefits of this technology:

• Streamlines data management processes
• Enhances data quality and consistency
• Increases efficiency and productivity
• Enables better decision-making based on linked records

Abstract

Records linking is provided. Two records are selected from a plurality of records corresponding to a customer for pair-wise record comparison. It is determined whether the two records are included in different entities. A local auto-link-threshold value of the different entities is identified in response to determining that the two records are included in different entities. An attribute comparison is performed between the two records. A comparison score is generated based on the attribute comparison between the two records. It is determined whether the comparison score is greater than the local auto-link-threshold value of the different entities. The two records are linked in response to determining that the comparison score is greater than the local auto-link-threshold value of the different entities.

GENERATIONAL ACCESS TO SAFEGUARDED COPY SOURCE VOLUMES (17809348)

Main Inventor

Theresa Mary Brown

Brief explanation

Ensuring recoverability and usability of potentially corrupted enterprise data by creating safeguarded copy volumes.

• Detecting data corruption and locating the last known uncorrupted copy quickly.
• Initiating restoration process with the identified uncorrupted data and reapplying subsequent logged transactions if possible.
• Bringing the data to the most recent and uncorrupted version.

Potential Applications

This technology can be applied in various industries and sectors where data integrity and recoverability are crucial, including:

• Banking and financial institutions
• Healthcare and medical records management
• Government agencies and public sector organizations
• E-commerce and online platforms
• Data centers and cloud service providers

Problems Solved

The technology addresses the following problems:

• Data corruption in enterprise systems
• Difficulty in detecting and locating the last known uncorrupted data
• Inability to restore data to the most recent and uncorrupted version
• Loss of critical information due to corruption

Benefits

The benefits of this technology include:

• Enhanced data integrity and recoverability
• Quick detection and location of uncorrupted data
• Efficient restoration process
• Minimized data loss and improved data reliability

Abstract

Ensuring that enterprise data that has potentially become corrupted is recoverable and usable by creating safeguarded copy volumes of the enterprise data. One important aspect of any corruption detection scheme is to determine when the data has become corrupted and locate the last known uncorrupted copy of the data as quickly as possible. Once this set of data is identified, the restoration process can begin with that data, and subsequent logged transactions can be reapplied if possible, which brings the data to the most recent and uncorrupted version.

FINE GRANULARITY READ ACCESS TO GENERATIONAL SAFEGUARDED COPY DATA (17809362)

Main Inventor

William J. Rooney

Brief explanation

The patent application describes a method for accessing specific copies of data without the need for restoration or compromising other copies. 

• The method allows for fine granularity read access to generational copies of safeguarded copy data.
• It uses a generation ID to identify the exact copy that contains the relevant enterprise data.
• The access is done without restoring the data to a recovery volume.
• It ensures the security and integrity of other copies of the data.

Potential Applications

This technology can be applied in various industries and scenarios, including:

• Data backup and recovery systems
• Disaster recovery planning
• Enterprise data management
• Cloud storage and retrieval systems

Problems Solved

The technology addresses the following problems:

• Time-consuming and resource-intensive data restoration processes
• Risk of compromising other copies of data during access
• Difficulty in identifying the specific copy containing relevant data

Benefits

The technology offers several benefits:

• Efficient and quick access to specific copies of data
• Reduced resource requirements for data recovery
• Enhanced data security and integrity
• Improved data management and retrieval processes

Abstract

Providing fine granularity read access to generational copies of safeguarded copy data. In some instances, a fine granularity read access to generational copies of safeguarded copy data includes the ability to use a generation ID to determine the exact generation copy that contains relevant enterprise related data that must be recovered and/or utilized. Additionally, the fine granularity read access is done in a manner that does not require a restore of the relevant enterprise related data to a recovery volume and that does not compromise other generation copies.

DATA PRIVACY WORKLOAD DISTRIBUTION IN A MULTI-TENANT HYBRID CLOUD COMPUTING ENVIRONMENT (17808233)

Main Inventor

Uwe Karl Hansmann

Brief explanation

The patent application describes a method to improve service routing by routing service requests to specific execution environments based on trust levels. Here are the key points:

• The system provides multiple execution environments where executable services can be deployed.
• A service registry is used to maintain information about these execution environments.
• When a service routing request is received, a trained machine-learning system is used to determine the required trust level for the service.
• The service registry then identifies a set of execution environments that match the trust level determined by the machine-learning system.
• From the set of execution environments, one is selected by the service registry.
• Finally, the service request is routed to the selected execution environment for execution.

Potential applications of this technology:

• Cloud computing platforms can use this method to efficiently route service requests to appropriate execution environments based on trust levels.
• Service-oriented architectures can benefit from improved service routing, ensuring that services are executed in trusted environments.
• Internet of Things (IoT) systems can use this method to route service requests to specific devices or environments based on trust levels.

Problems solved by this technology:

• Efficient service routing: The method ensures that service requests are routed to appropriate execution environments based on trust levels, improving overall system efficiency.
• Trust-based execution: By considering trust levels, the method ensures that services are executed in environments that meet the required level of trust, enhancing security and reliability.

Benefits of this technology:

• Improved system efficiency: By routing service requests to appropriate execution environments, the method optimizes resource utilization and reduces response times.
• Enhanced security and reliability: By considering trust levels, the method ensures that services are executed in environments that meet the required level of trust, reducing the risk of security breaches and improving overall system reliability.

Abstract

In an approach to improve service routing, embodiments route a service request to an execution environment. Embodiments provide a plurality of execution environments, wherein in each execution environment executable services are deployable, provide a service registry maintaining a plurality of execution environments, and receive, by the service registry, a service routing request. Further, embodiments determine a required trust level for a service relating to the service routing request by using a trained machine-learning system for outputting a trust level class when receiving service context data of the service relating to the service routing request as input, determine, using the service registry, a set of execution environments matching the output trust level class, and select, by the service registry, one execution environment of the determined set of execution environments. Further, embodiments route, by the service registry, the service request to the selected one of the execution environments for execution.

PROACTIVE AUDIBLE SOUND REVERBERATION MITIGATION FOR PREDICTED USER EXPERIENCE (17809295)

Main Inventor

Tiberiu Suto

Brief explanation

Abstract:

This patent application describes a method for using digital twin simulation to automatically reduce noise in an environment. The method involves creating a digital twin of the environment, simulating vibration in the environment based on the activities and equipment within the digital twin, determining how the vibration propagates within the environment, and generating a plan to mitigate the vibration for a user in the environment.

Patent/Innovation Explanation:

• A digital twin of an environment is created.
• The digital twin simulates vibration in the environment based on the activities and equipment within the digital twin.
• The propagation of vibration within the environment is determined based on the simulated vibration generated by the digital twin.
• A plan is generated to mitigate the vibration for a user in the environment.

Potential Applications:

• Noise reduction in industrial settings.
• Noise reduction in residential areas.
• Noise reduction in transportation systems.
• Noise reduction in healthcare facilities.

Problems Solved:

• Reducing noise pollution in various environments.
• Improving the comfort and well-being of individuals in noisy environments.
• Enhancing the efficiency and productivity of workers in noisy industrial settings.

Benefits:

• Automatic noise mitigation without the need for manual intervention.
• Real-time simulation and analysis of vibration and noise.
• Customized plans for noise reduction tailored to specific environments.
• Improved user experience and satisfaction in noisy environments.
• Increased productivity and efficiency in workplaces.

Abstract

In an approach for utilizing digital twin simulation for automatically mitigating noise, a processor generates a digital twin of an environment. The digital twin simulates vibration within the environment based on equipment and activities within the environment that are simulated by the digital twin. A processor determines how vibration is propagated within the environment based on the simulated vibration generated by the digital twin. A processor generates a plan for mitigating the vibration for a user within the environment.

CONTEXT AWARE OPTIMIZED DIGITAL TWIN MODEL CREATION (17809066)

Main Inventor

Shailendra Moyal

Brief explanation

The patent application describes a method for building a subset of a digital twin model for a machine. A digital twin model is a virtual representation of a physical machine that can be used for monitoring and analysis purposes.

• The processor monitors the digital twin model of the machine for any changes.
• If a change is detected, the processor analyzes the contextual situation of the machine and the activities performed by the machine.
• If the contextual situation has changed, the processor analyzes the data generated by the sensors of the machine.
• The processor determines if the sensors are generating the required type and amount of data to create a subset of the digital twin model.
• If the requirements are met, the processor creates the subset of the digital twin model incorporating the detected change.

Potential Applications

• Industrial machinery monitoring and analysis
• Predictive maintenance for machines
• Optimization of machine performance

Problems Solved

• Efficiently updating a digital twin model when changes occur in the machine
• Ensuring that the subset of the digital twin model includes the necessary data for accurate analysis

Benefits

• Real-time monitoring and analysis of machine performance
• Improved predictive maintenance capabilities
• Enhanced optimization of machine operations

Abstract

In an approach for building a subset of a digital twin model, a processor monitors a digital twin model of a machine for one or more changes to the machine. Responsive to detecting a change to the machine, a processor analyzes one or more aspects of the machine, wherein the one or more aspects of the machine include a contextual situation of the machine and one or more activities performed by the machine. Responsive to determining the contextual situation of the machine has changed, a processor analyzes one or more sensors and a first set of data generated by the one or more sensors. A processor determines the one or more sensors are generating a required type of data and a required amount of data to create a subset of the digital twin model. A processor creates the subset of the digital twin model incorporating the change detected.

ESTIMATING EMISSION SOURCE LOCATION FROM SATELLITE IMAGERY (17809037)

Main Inventor

Theodore G. van Kessel

Brief explanation

The patent application describes a method for estimating the location of emission sources using satellite plume data. Here are the key points:

• The method involves creating a dataset of plume concentration data from satellite observations.
• The dataset is then downsampled to match the resolution of the satellite.
• The downscaled dataset is divided into two separate datasets based on a predetermined proportion.
• Two machine learning models are trained using at least one of the two datasets.
• The first model is designed to identify the presence of a plume, while the second model identifies the position and magnitude of the plume source.
• The trained models are then applied to new concentration data to estimate the emission source location.

Potential applications of this technology:

• Environmental monitoring: The method can be used to track and identify emission sources, helping in the monitoring and regulation of air pollution.
• Industrial emissions control: Industries can use this method to locate and quantify their emission sources, enabling them to take necessary measures to reduce pollution.
• Emergency response: The method can aid in quickly identifying the source of hazardous emissions during emergencies, allowing for prompt action to protect public health.

Problems solved by this technology:

• Traditional methods of estimating emission source locations are often time-consuming and require extensive manual analysis.
• This method automates the process using machine learning, making it more efficient and accurate.
• It provides a scalable solution for analyzing large amounts of satellite plume data.

Benefits of this technology:

• Improved accuracy: The use of machine learning models enhances the accuracy of estimating emission source locations.
• Time and cost savings: The automated process reduces the need for manual analysis, saving time and resources.
• Scalability: The method can handle large datasets, making it suitable for analyzing satellite plume data on a global scale.

Abstract

In an approach for estimating emission source location from satellite plume data, a processor creates a dataset of plume concentration data. A processor down samples the dataset to an array at satellite resolution. A processor partitions the array into two separate datasets according to a preset proportion. A processor trains two machine learning models on at least one of the two separate datasets, wherein a first machine learning model of the two machine learning models is for identifying a presence of a plume and a second machine learning model of the two machine learning models is for identifying a source position and magnitude of the plume. A processor applies the two machine learning models to new concentration data.

GENERATING GOAL-ORIENTED DIALOGUES FROM DOCUMENTS (17808628)

Main Inventor

Song Feng

Brief explanation

The patent application describes a computer-based method, system, and program for generating a goal-oriented dialogue from a grounding document. This involves analyzing a corpus of text to identify semantic structures that can be used to simulate a dialogue. The system then generates a simulated dialogue flow, including utterances from a simulated agent and user.

• The technology analyzes a body of text to identify structures that can be used to simulate a dialogue.
• It generates a simulated dialogue flow, including utterances from a simulated agent and user.
• The goal is to create a goal-oriented dialogue based on the content of a grounding document.

Potential Applications

• Virtual assistants or chatbots that can engage in goal-oriented conversations with users.
• Customer service systems that can provide personalized and interactive support.
• Educational platforms that can simulate dialogue-based learning experiences.

Problems Solved

• Automates the process of generating goal-oriented dialogues from a grounding document.
• Enables the creation of interactive and personalized conversational systems.
• Improves the efficiency and effectiveness of virtual assistants and chatbots.

Benefits

• Enhances user experience by providing interactive and engaging conversations.
• Saves time and resources by automating the dialogue generation process.
• Enables the development of more advanced and intelligent conversational systems.

Abstract

Provided is a computer-implemented method, system, and computer program product for generating a goal-oriented dialogue from a grounding document. A processor may analyze a corpus of text. The processor may identify, based on the analyzing, one or more semantic structures that can be used to simulate a dialogue. The processor may generate, based on the identifying, a simulated dialogue, the simulated dialogue including one or more utterances from a simulated agent and one or more utterances from a simulated user to form a dialogue flow.

DYNAMIC MEETING ATTENDEE INTRODUCTION GENERATION AND PRESENTATION (17808166)

Main Inventor

Jacob Ryan Jepperson

Brief explanation

The patent application describes a system for generating dynamic introductions in virtual meetings by analyzing user data and using natural language processing. 

• The system gathers introduction data of a user in a virtual meeting.
• The gathered data is analyzed using natural language processing to identify user features.
• User features are captured to create a dynamic introduction statement based on the scheduled meeting.
• The dynamic introduction statement is presented and user features are captured to update it during a live meeting.

Potential Applications

• Virtual meetings and conferences
• Online networking events
• Webinars and online training sessions

Problems Solved

• Time-consuming and repetitive introductions in virtual meetings
• Lack of personalization and engagement in virtual meetings

Benefits

• Saves time by automatically generating introductions
• Enhances personalization and engagement in virtual meetings
• Improves the overall experience of virtual meetings

Abstract

The embodiment may include dynamic introduction generation for a virtual meeting that may gather introduction data of a user. The embodiment may analyze the gathered introduction data using natural language processing to identify user features. The embodiment may capture the user features for a dynamic introduction statement based on determining a scheduled meeting. The embodiment may present the dynamic introduction statement and capture user features to update the dynamic introduction statement based on determining a live meeting.

REVERSE REINFORCEMENT LEARNING TO TRAIN TRAINING DATA FOR NATURAL LANGUAGE PROCESSING NEURAL NETWORK (17808511)

Main Inventor

Zhong Fang Yuan

Brief explanation

The abstract describes a computer-implemented process for modifying a training dataset using a neural network and reverse reinforcement learning. Here is a simplified explanation of the abstract:

• The process starts by benchmarking the training dataset using a State Of The Art (SOTA) neural network to establish a performance benchmark.
• The training dataset is then divided into smaller slices.
• A selection strategy generator is used to choose a sequence of atomic operations to be applied to one of the slices.
• The selected sequence of operations modifies the slice, creating a revised version.
• Reverse reinforcement learning is performed on the revised slice using the benchmark and the SOTA neural network.
• Finally, the original slice is replaced with the revised slice in the training dataset, resulting in a modified training dataset.

Potential applications of this technology:

• Improving the performance of machine learning models by fine-tuning the training dataset.
• Enhancing the accuracy and efficiency of neural networks in various domains, such as image recognition, natural language processing, and recommendation systems.

Problems solved by this technology:

• Addressing the limitations of traditional training dataset modification methods by using a combination of benchmarking, atomic operations, and reverse reinforcement learning.
• Overcoming the challenges of manually modifying large training datasets by automating the process.

Benefits of this technology:

• Increased accuracy and performance of machine learning models by optimizing the training dataset.
• Time and cost savings by automating the dataset modification process.
• Improved generalization and robustness of neural networks by incorporating reverse reinforcement learning.

Abstract

A computer-implemented process for modifying a training dataset includes the following operations. The training dataset is benchmarked using a State Of The Art (SOTA) neural network to determine a benchmark for the training dataset. The training set is divided into a plurality of slices. A sequence of a plurality of atomic operations are selected using a selection strategy generator operating on one of the plurality of slices. The sequence of the plurality of atomic operations is applied to modify the one of the plurality of slices to generate a revised one of the plurality of slices. Reverse reinforcement learning is performed on the revised one of the plurality of slices using the benchmark and the SOTA neural network. The training dataset is modified by replacing the one of the plurality of slices with the revised one of the plurality of slices to generate a modified training dataset.

BUNDLING HYPERVECTORS (17809052)

Main Inventor

Michael Andreas Hersche

Brief explanation

The patent application describes a method for bundling a set of code hypervectors using weights and mapping techniques. Here are the key points:

• The method involves bundling a set of M code hypervectors, each with a dimension of D, where M is greater than 1.
• A weight vector is received, which is an M-dimensional vector containing weights for weighting the set of code hypervectors.
• The weight vector is then mapped to an S-dimensional vector, s, where each element of s indicates one of the code hypervectors. The mapping is such that S is equal to D divided by L, and L is greater than or equal to 1.
• Finally, a hypervector is built using the mapped vector, where each element of the built hypervector corresponds to the element of the code hypervector indicated in the corresponding element of the S-dimensional vector, s.

Potential applications of this technology:

• Pattern recognition: The method can be used to bundle code hypervectors representing patterns or features, allowing for efficient recognition and classification of similar patterns.
• Data compression: By bundling multiple code hypervectors into a single hypervector, the method can be used for data compression, reducing storage requirements.
• Machine learning: The technique can be applied in machine learning algorithms to improve the efficiency and accuracy of training and inference processes.

Problems solved by this technology:

• Efficient representation: The method provides a way to bundle multiple code hypervectors into a single hypervector, reducing the complexity and storage requirements of representing multiple patterns or features.
• Weighted bundling: The use of weights allows for prioritization or emphasis on specific code hypervectors, enabling more flexible and customizable bundling.
• Mapping technique: The mapping of the weight vector to the S-dimensional vector provides a systematic way to indicate which code hypervectors are included in the bundled hypervector.

Benefits of this technology:

• Improved efficiency: The bundling of code hypervectors reduces the computational complexity and storage requirements, leading to faster processing and reduced memory usage.
• Customizability: The use of weights and mapping techniques allows for customization and prioritization of code hypervectors, enabling tailored solutions for specific applications.
• Versatility: The method can be applied in various domains such as pattern recognition, data compression, and machine learning, making it a versatile solution for different use cases.

Abstract

Embodiments are disclosed for a method. The method includes bundling a set of M code hypervectors, each of dimension D, where M>1. The bundling includes receiving an M-dimensional vector comprising weights for weighting the set of code hypervectors. The bundling further includes mapping the M-dimensional vector to an S-dimensional vector, s, such that each element of the S-dimensional vector, s, indicates one of the set of code hypervectors, where S=D/L and L≥1. Additionally, the bundling includes building a hypervector such that an ith element of the built hypervector is an ith element of the code hypervector indicated in an ith element of the S-dimensional vector, s.

BLOCKWISE FACTORIZATION OF HYPERVECTORS (17809044)

Main Inventor

Michael Andreas Hersche

Brief explanation

The abstract describes a method for determining the cognitive concepts represented by a data structure using hypervectors. Here are the key points:

• The method determines the granularity of hypervectors, which are mathematical representations of data structures.
• It receives an input hypervector that represents a data structure.
• The method performs an iterative process to break down the input hypervector into individual hypervectors that represent cognitive concepts.
• For each concept, an unbound version of a hypervector representing the concept is determined by comparing it with estimate hypervectors of other concepts.
• A similarity vector is generated to indicate the similarity between the unbound version of the hypervector and each candidate code hypervector of the concept.
• An estimate of a hypervector representing the concept is generated by combining the candidate code hypervectors and the weights of the similarity vector.

Potential applications of this technology:

• Cognitive computing: This method can be used to analyze and understand complex data structures, such as natural language processing or image recognition.
• Artificial intelligence: The method can help in building intelligent systems that can learn and recognize patterns in data.
• Data analysis: By breaking down data structures into cognitive concepts, this method can assist in analyzing and extracting meaningful insights from large datasets.

Problems solved by this technology:

• Complex data analysis: The method provides a systematic approach to break down and analyze complex data structures, making it easier to understand and extract useful information.
• Pattern recognition: By representing data structures as hypervectors and breaking them down into cognitive concepts, the method enables the recognition of patterns and similarities in the data.

Benefits of this technology:

• Improved data understanding: The method helps in understanding the underlying cognitive concepts represented by data structures, leading to better insights and decision-making.
• Efficient data analysis: By factorizing the input hypervector into individual hypervectors, the method simplifies the analysis process and reduces computational complexity.
• Scalability: The method can be applied to large datasets and can handle a wide range of data structures, making it scalable for various applications.

Abstract

Embodiments are disclosed for a method. The method includes determining a granularity of hypervectors. The method also includes receiving an input hypervector representing a data structure. Additionally, the method includes performing an iterative process to factorize the input hypervector into individual hypervectors representing the cognitive concepts. The iterative process includes, for each concept: determining an unbound version of a hypervector representing the concept by a blockwise unbinding operation between the input hypervector and estimate hypervectors of other concepts. The iterative process further includes determining a similarity vector indicating a similarity of the unbound version of the hypervector with each candidate code hypervector of the concept. Additionally, the iterative process includes generating an estimate of a hypervector representing the concept by a linear combination of the candidate code hypervectors, and weights of the similarity vector.

CROSSBAR ARRAYS IMPLEMENTING TRUTH TABLES (17847955)

Main Inventor

Charles Mackin

Brief explanation

The abstract describes a method for preparing a trained crossbar array of a neural network. Here is a simplified explanation of the abstract:

• The method involves using a computer simulation of a crossbar array to train the neural network.
• A predetermined truth table is used as input, and analog output values are generated based on simulated weights.
• Loss values are calculated by comparing the analog output values with expected values for the output portion of the truth table.
• The simulated weights are adjusted based on the calculated loss values.
• The input portion of the truth table is repeatedly fed into the simulation and the output values are recalculated using the adjusted weights.
• This process continues until the analog output values match the expected values within a predefined margin of error.

Potential applications of this technology:

• Artificial intelligence and machine learning systems
• Pattern recognition and image processing
• Natural language processing and speech recognition
• Autonomous vehicles and robotics
• Financial analysis and prediction

Problems solved by this technology:

• Efficient training of neural networks using a computer simulation
• Optimization of weights in a crossbar array for accurate output prediction
• Reducing the margin of error in neural network predictions

Benefits of this technology:

• Faster and more efficient training of neural networks
• Improved accuracy and reliability of neural network predictions
• Cost-effective implementation of neural network systems
• Scalability for handling large datasets and complex problems

Abstract

A method for preparing a trained crossbar array of a neural network is provided. The method includes feeding an input portion of a predetermined truth table into a computer simulation of a crossbar array, and generating analog output values for the input portion of the truth table based on simulated weights. The method further includes calculating a loss value from each of the analog output values and expected values for an output portion of the truth table, and adjusting the simulated weights based on the calculated loss values. The method further includes refeeding the input portion of the predetermined truth table into the computer simulation and recalculating the output values using the adjusted simulated weights until the analog output values produce the expected values for the output portion of the truth table within a predefined margin of error.

Stochastic Bitstream Generation with In-Situ Function Mapping (17848136)

Main Inventor

Pritish Narayanan

Brief explanation

The patent application describes techniques for generating digital outputs using stochastic bitstreams and activation function mapping. The system includes a shared circuitry component with a random number generator (RNG) and a digital-to-analog converter (DAC). 

• The RNG generates a sequence of random addresses to read a random sequence of digital voltage references stored in a lookup table (LUT).
• The DAC converts the random sequence of digital voltage references into random analog voltage references (V).
• The system also includes comparators that compare the random analog voltage references (V) with input analog voltages (Vin) to produce sequences of digital pulses as stochastic bitstreams.
• Multiple comparators can be used to simultaneously compare each random analog voltage reference (V) against more than one input analog voltage (Vin) in parallel.

Potential applications of this technology:

• Random number generation for cryptography and security applications.
• Simulation and modeling in scientific research and engineering.
• Neural networks and machine learning algorithms.
• Signal processing and communication systems.

Problems solved by this technology:

• Generating random bitstreams for various applications.
• Efficiently converting digital voltage references to analog voltage references.
• Simultaneously comparing multiple analog voltage references with input analog voltages.

Benefits of this technology:

• Provides a reliable and efficient method for generating random bitstreams.
• Enables parallel comparison of multiple analog voltage references.
• Can be used in a wide range of applications, including cryptography, simulation, and machine learning.

Abstract

Techniques for generating digital outputs as stochastic bitstreams with activation function mapping are provided. In one aspect, a system includes: a shared circuitry component including a RNG for generating a sequence of random addresses to read a random sequence of digital voltage references stored in a LUT, and a DAC for converting the random sequence of digital voltage references into random analog voltage references V; and a comparator(s) for comparing the random analog voltage references Vand input analog voltages Vin sequences of comparisons to produce sequences of digital pulses as stochastic bitstreams. A system having multiple comparators for simultaneously comparing each of the random analog voltage references Vagainst more than one of the input analog voltages Vin parallel is also provided, as is a method for generating digital outputs from input analog voltages V.

UNDERSTANDING REINFORCEMENT LEARNING POLICIES BY IDENTIFYING STRATEGIC STATES (17808181)

Main Inventor

Ronny Luss

Brief explanation

The patent application describes a method for generating explanations for a deep reinforcement learning policy.

• The method involves computing a maximum likelihood path matrix, which represents the shortest path between each state in a set of states associated with a trained model.
• Explanations are generated based on identified meta-states and selected strategic states, using the computed maximum likelihood path matrix.

Potential Applications

This technology has potential applications in various fields, including:

• Artificial intelligence
• Machine learning
• Reinforcement learning
• Robotics
• Autonomous systems

Problems Solved

The technology addresses the following problems:

• Lack of interpretability in deep reinforcement learning policies
• Difficulty in understanding the decision-making process of AI systems
• Limited ability to explain the reasoning behind AI-driven actions

Benefits

The technology offers several benefits, such as:

• Improved transparency and interpretability of deep reinforcement learning policies
• Enhanced understanding of AI system decision-making
• Ability to provide explanations for AI-driven actions
• Facilitation of trust and accountability in AI systems

Abstract

One or more computer processors compute a maximum likelihood path matrix comprising a respective shortest path between each state in a set of states associated with a model trained with a deep reinforcement learning policy. The one or more computer processors generate explanations for the deep reinforcement learning policy based one or more identified meta-states for each state in the set of states and corresponding selected strategic states utilizing the computed maximum likelihood path matrix.

DYNAMIC RESOURCE ALLOCATION METHOD FOR SENSOR-BASED NEURAL NETWORKS USING SHARED CONFIDENCE INTERVALS (17809310)

Main Inventor

Paul Schardt

Brief explanation

The abstract describes a method, computer program, and computer system for resource allocation in sensor-based neural networks. Here is a simplified explanation of the abstract:

• The invention provides a method, computer program, and computer system for allocating resources in sensor-based neural networks.
• The method involves identifying one or more nodes in an edge computing environment.
• Data from these nodes, which includes a classification dataset with reference classification and confidence value data, is received.
• A node is selected from the identified nodes based on having the highest confidence interval associated with the reference classification.
• The selected node is then assigned to process the classification dataset.

Potential Applications

This technology has potential applications in various fields, including:

• Internet of Things (IoT) systems
• Edge computing environments
• Sensor networks
• Artificial intelligence and machine learning systems

Problems Solved

The technology addresses the following problems:

• Efficient allocation of resources in sensor-based neural networks
• Optimizing the processing of classification datasets in edge computing environments
• Improving the accuracy and reliability of classification tasks in distributed systems

Benefits

The technology offers several benefits, including:

• Improved resource allocation and utilization in sensor-based neural networks
• Enhanced accuracy and efficiency in processing classification datasets
• Increased reliability and performance of edge computing environments
• Facilitates distributed processing and decision-making in IoT systems

Abstract

A method, computer program, and computer system are provided for resource allocation for sensor-based neural networks. One or more nodes associated with an edge computing environment are identified. Data corresponding to a classification dataset is received from the identified nodes. The dataset includes a reference classification and confidence value data. A node is selected from among the identified nodes based on the selected node having a greatest confidence interval associated with the reference classification within the confidence value data. The selected node is assigned to process the classification dataset.

MULTIPLE STAGE KNOWLEDGE TRANSFER (17849969)

Main Inventor

Amit Dhurandhar

Brief explanation

Abstract:

An input model is received along with a set of requirements, which describe an output model to be trained. The output model is then trained based on the input model and at least one intermediate model.

Patent/Innovation Explanation:

• An input model is provided along with a set of requirements.
• The requirements specify the desired output model that needs to be trained.
• The output model is trained using the input model as a basis.
• The training process also incorporates at least one intermediate model to further enhance the output model.

Potential Applications:

• Machine learning and artificial intelligence systems.
• Data analysis and prediction models.
• Image recognition and computer vision algorithms.
• Natural language processing and speech recognition systems.

Problems Solved:

• Efficiently training complex models by utilizing intermediate models.
• Meeting specific requirements for the output model.
• Improving the accuracy and performance of trained models.
• Streamlining the training process for various applications.

Benefits:

• Enhanced accuracy and performance of trained models.
• Flexibility to meet specific requirements and desired output.
• Improved efficiency in training complex models.
• Potential for advancements in various fields such as AI, data analysis, and image recognition.

Abstract

An input model can be received, along with a set of requirements. The set of requirements may describe an output model to be trained. The output model can then be trained. The training of the output model can be based on the input model and based further on at least one intermediate model.

SELF-LEARNED REFERENCE MECHANISM FOR COMPUTER MODEL SELECTION (17809354)

Main Inventor

Jiang Bo Kang

Brief explanation

The patent application describes a method, system, and computer program for self-learning reference mechanisms in AutoAI model selection. Here is a simplified explanation of the abstract:

• The method identifies key statistics within a dataset.
• It groups these statistics based on data patterns.
• The method determines if a data pattern group is mature.
• If a group is mature, a model selection acceleration mechanism (MSAM) model is generated.
• The MSAM model is then used to predict the top-k models for the dataset.

Potential applications of this technology:

• Automated machine learning (AutoML) platforms
• Data analysis and decision-making systems
• Predictive modeling and forecasting tools

Problems solved by this technology:

• Manual model selection can be time-consuming and subjective.
• Identifying the most suitable models for a dataset can be challenging.
• Keeping up with the constantly evolving field of machine learning can be difficult.

Benefits of this technology:

• Saves time and effort by automating the model selection process.
• Improves accuracy and reliability by using self-learning reference mechanisms.
• Enables faster and more efficient decision-making based on predictive models.

Abstract

A method, system, and computer program product for self-learning reference mechanisms for model selection in AutoAI. The method identifies a set of data summary statistics within a data set. A data pattern group is identified within the set of data summary statistics. The data pattern group is determined to be mature. A model selection acceleration mechanism (MSAM) model is generated based on the data pattern group. The method predicts a set of top-k models for the data set based on the MSAM model.

BLACK-BOX EXPLAINER FOR TIME SERIES FORECASTING (17808314)

Main Inventor

Vikas C. Raykar

Brief explanation

The patent application describes a method, system, and computer program for explaining the predictions of univariate time series forecasters that use black-box models. Here are the key points:

• The method receives a set of time series forecasting predictions generated by black-box models trained with an initial data set.
• A set of features is generated based on at least a portion of the initial data set.
• Surrogate models are trained using the set of time series forecasting predictions and the set of features.
• A subset of surrogate models is selected.
• Based on the subset of surrogate models, one or more explanation outputs are generated for the time series forecasting predictions of the black-box models.

Potential applications of this technology:

• Interpreting the predictions of univariate time series forecasters.
• Providing explanations for the predictions made by black-box models.
• Enhancing transparency and trust in the decision-making process of time series forecasting models.

Problems solved by this technology:

• Black-box models often lack interpretability, making it difficult to understand the reasoning behind their predictions.
• Explaining the predictions of time series forecasters can help identify potential biases, errors, or anomalies in the models.
• Providing explanations can improve the adoption and acceptance of time series forecasting models in various domains.

Benefits of this technology:

• Enables users to understand and trust the predictions made by black-box time series forecasters.
• Facilitates the identification of potential issues or limitations in the forecasting models.
• Enhances transparency and accountability in decision-making processes that rely on time series forecasting.

Abstract

A method, system, and computer program product for an interpretable, feature-based post-hoc black box explainer for univariate time series forecasters are provided. The method receives a set of time series forecasting predictions. The set of time series forecasting predictions are generated from a set of black-box models trained with an initial data set. The method generates a set of features based on at least a portion of the initial data set. A set of surrogate models are trained based on the set of time series forecasting predictions and at least a portion of the set of features. A subset of surrogate models is selected. Based on the subset of surrogate models, the method generates one or more explanation outputs for time series forecasting predictions of the set of black-box models.

GLOBAL CONTEXT EXPLAINERS FOR ARTIFICIAL INTELLIGENCE (AI) SYSTEMS USING MULTIVARIATE TIMESERIES DATA (17808927)

Main Inventor

Vijay ARYA

Brief explanation

The patent application describes techniques for explaining the behavior of Artificial Intelligence systems that use multivariate timeseries data. Here are the key points:

• Predictions are made for multivariate timeseries data.
• Feature importance weights are generated using a feature-based local explainer.
• Each feature importance weight is associated with a time period and a corresponding data source of the timeseries data.
• A dataset is created using the feature importance weights, with labels indicating whether the weight is positive or negative for each time period and data source.
• One or more global explanations are generated using the dataset and a rule-based explainer.
• The global explanations show how the predictions change at specific times in the timeseries data based on values from the corresponding data source.
• An action can be taken based on the global explanations.

Potential applications of this technology:

• Understanding the behavior of AI systems that use multivariate timeseries data.
• Identifying important features and their impact on predictions.
• Debugging and improving the performance of AI systems.
• Enhancing transparency and interpretability of AI systems.

Problems solved by this technology:

• Lack of transparency in AI systems using multivariate timeseries data.
• Difficulty in understanding the factors influencing predictions.
• Limited interpretability of AI systems using complex data.
• Challenges in debugging and improving the performance of AI systems.

Benefits of this technology:

• Provides explanations for the behavior of AI systems using multivariate timeseries data.
• Helps identify important features and their impact on predictions.
• Enhances transparency and interpretability of AI systems.
• Facilitates debugging and improvement of AI system performance.

Abstract

Provided are techniques for global context explainers for Artificial Intelligence systems using multivariate timeseries data. Predictions for multivariate timeseries data are received. Feature importance weights are generated from the predictions using a feature-based local explainer, where each of the feature importance weights is associated with a time period and a corresponding data source of timeseries data of the multivariate timeseries data. A dataset is generated using the feature importance weights, where the dataset includes, for each time period and the corresponding data source, a label indicating whether the feature importance weight is one of positive and negative. One or more global explanations are generated using the dataset and a directly interpretable rule-based explainer, where the one or more global explanations indicate how the predictions change at particular times in the multivariate timeseries data based on values from the corresponding data source. An action based on the global explanations is performed.

QUANTUM ADVANTAGE USING QUANTUM CIRCUIT FOR GRADIENT ESTIMATION (17990473)

Main Inventor

Nikitas Stamatopoulos

Brief explanation

The patent application describes quantum gradient algorithms that offer an advantage over conventional methods in quantum computing systems. These algorithms are implemented using a quantum circuit and involve a phase oracle O based on a finite difference approximation with an order greater than zero. The complexity of the algorithm scales as (√{square root over (k)}/ϵ), where k represents the dimensionality of the gradient and ϵ is the desired error tolerance.

• Quantum gradient algorithms provide a quantum advantage over conventional methods.
• The algorithms are implemented using a quantum circuit on qubits of a quantum computing system.
• A phase oracle O is used, which is based on a finite difference approximation with an order greater than zero.
• The complexity of the algorithm scales as (√{square root over (k)}/ϵ), where k is the dimensionality of the gradient and ϵ is the desired error tolerance.
• The quantum circuit is repeatedly executed to determine a k-dimensional gradient of a function ƒ(x) within an error ϵ at point x.

Potential Applications

• Optimization problems in various fields such as finance, logistics, and engineering can benefit from the improved efficiency of quantum gradient algorithms.
• Machine learning and artificial intelligence applications that involve gradient-based optimization can be enhanced using these algorithms.
• Quantum chemistry simulations and molecular modeling can be improved by leveraging the quantum advantage provided by these algorithms.

Problems Solved

• Conventional methods for gradient-based optimization can be computationally expensive and time-consuming, especially for high-dimensional problems.
• Quantum gradient algorithms offer a more efficient approach to compute gradients, reducing the computational burden and enabling faster optimization.
• These algorithms address the challenges faced in quantum computing systems by utilizing the unique properties of quantum circuits and qubits.

Benefits

• Quantum gradient algorithms provide a quantum advantage, enabling faster and more efficient optimization compared to conventional methods.
• The use of quantum circuits and qubits allows for parallel computation, leading to significant speedup in gradient calculations.
• By reducing the computational complexity, these algorithms pave the way for solving larger and more complex optimization problems in various domains.

Abstract

Described herein are quantum gradient algorithms that result in a quantum advantage over conventional methods. In an example, a quantum circuit is configured to implement a quantum gradient algorithm when executed on qubits of a quantum computing system. The quantum gradient algorithm includes a phase oracle Odefined by a finite difference approximation with an order greater than zero, and a complexity of the quantum gradient algorithm scales as (√{square root over (k)}/ϵ). The quantum circuit is repeatedly executed on qubits of a quantum computing system to determine a k-dimensional gradient of a function ƒ(x) within an error ϵ at point x.

COMPOSING A MACHINE LEARNING MODEL FOR COMPLEX DATA SOURCES (17808143)

Main Inventor

Ana Paula Appel

Brief explanation

The patent application describes a method for creating a machine learning model for complex data sources. Here are the key points:

• The computer receives data and metadata related to a machine learning task from a user.
• It determines the task context and problem domain.
• The computer identifies the specific machine learning task.
• It evaluates pre-compiled models to find a match with the problem domain.
• At least two pre-compiled models are selected.
• The computer generates multiple combinations of these models.
• The combinations are executed with the data and metadata.
• The results of the executed combinations are displayed to the user.
• The computer checks if the user finds the level of error in the results acceptable.

Potential Applications

This technology has potential applications in various fields, including:

• Healthcare: Creating machine learning models to analyze medical data and assist in diagnosis.
• Finance: Developing models to predict market trends and make investment decisions.
• Manufacturing: Optimizing production processes by analyzing complex data from sensors and machines.
• Natural Language Processing: Building models to understand and generate human-like text.

Problems Solved

This technology addresses the following problems:

• Complex data sources: It enables the creation of machine learning models for data that is difficult to analyze using traditional methods.
• Model selection: It automates the process of selecting and combining pre-compiled models, saving time and effort.
• Error evaluation: It allows users to determine if the level of error in the results is acceptable, ensuring the reliability of the models.

Benefits

The use of this technology offers several benefits:

• Improved accuracy: By combining multiple models, the accuracy of predictions and analysis can be enhanced.
• Time and resource savings: The automated model selection process saves time and effort compared to manual selection.
• User feedback: The ability to evaluate the acceptability of error levels based on user response ensures user satisfaction and trust in the models.

Abstract

In an approach to composing a machine learning model for complex data sources, a computer receives data and associated metadata corresponding to a machine learning task from a user. A computer determines a task context and a problem domain. A computer identifies the machine learning task. A computer evaluates a match between the problem domain and one or more pre-compiled models. A computer selects at least two of the one or more pre-compiled models. A computer generates one or more multimodal model combinations with the selected at least two of the one or more pre-compiled models. A computer executes the multimodal model combinations with the data and associated metadata. A computer displays the results of the executed one or more multimodal model combinations to the user. A computer determines whether a level of error associated with the results is acceptable to the user based on a response from the user.

COMPUTER TRAINING DATA USING MACHINE LEARNING (17809463)

Main Inventor

Zhong Fang Yuan

Brief explanation

The abstract describes a method for training data models using machine learning. Here is a simplified explanation of the abstract:

• The method involves training a computer data model using a training data set.
• The training data set includes both training data and additional training data.
• The data distribution of the training data set is represented by layers of data.
• The computer data model is iteratively trained for each layer of the training data set.
• Statistical noise is randomly added to each layer of the training data set.
• Data variations are detected in each layer of the additional training data.
• The data variations are diluted in each additional layer of the training data.
• The computer data model is retrained using the diluted data variations in each layer of the additional training data.

Potential applications of this technology:

• Improving the accuracy and performance of machine learning models.
• Enhancing data analysis and prediction capabilities in various industries.
• Optimizing decision-making processes based on large datasets.

Problems solved by this technology:

• Overfitting: By adding statistical noise and diluting data variations, the method helps prevent overfitting, where a model becomes too specialized to the training data and performs poorly on new data.
• Data distribution representation: The use of layers to represent the data distribution allows for a more comprehensive understanding of the training data set.

Benefits of this technology:

• Improved model accuracy: By iteratively training the model for each layer and detecting data variations, the method helps create a more accurate and robust data model.
• Generalization capability: Diluting data variations helps the model generalize better to new data, improving its performance in real-world scenarios.
• Efficient training process: The iterative training approach allows for efficient utilization of the training data set, leading to faster and more effective model training.

Abstract

Training data models using machine learning can include training a computer data model of data distribution using a training data set. The training data set includes training data and additional training data, and the training data and the additional training data being represented by layers of data representing the data distribution of the training data set. The computer data model using the additional training data is iteratively trained for each of the layers of the training data set. Statistical noise is added randomly to each of the layers of the training data set. Data variations are detected in each of the layers of the additional training data. The data variations are diluted in each of the additional layers of the training data, and the computer data model is retrained for the training data set using the diluted data variations in each of the layers of the additional training data.

IDENTIFYING SKILL GAPS IN PROJECT TEAMS AND MATCHING WITH AVAILABLE RESOURCES (17809070)

Main Inventor

Madhusmita Patil

Brief explanation

The patent application describes a method for detecting resource gaps in a project and optimizing resource utilization to fill those gaps. Here are the key points:

• A processor monitors the progress of one or more projects on a project dashboard.
• If a project is found to have a resource gap, the processor collects data about the project.
• The processor predicts the productivity trajectory of the project.
• If a team member is predicted to fall short of the productivity trajectory, the processor identifies alternative combinations of team members who can meet the trajectory.
• The processor calculates the level of amplification needed to achieve the trajectory with the alternative team combinations.
• The processor identifies available resources to complete any pending tasks in the project backlog.
• The processor determines the best variation of resources to optimize project completion.

Potential applications of this technology:

• Project management software and tools can incorporate this method to help project managers identify and address resource gaps in real-time.
• Companies and organizations can use this method to optimize resource allocation and improve project efficiency.

Problems solved by this technology:

• Resource gaps in projects can lead to delays and inefficiencies. This method helps detect and address these gaps promptly.
• It provides a solution for predicting and managing team member productivity to ensure project success.

Benefits of this technology:

• Improved project efficiency and timely completion by addressing resource gaps.
• Optimal utilization of available resources to maximize productivity.
• Real-time monitoring and decision-making to adapt to changing project requirements.

Abstract

In an approach for detecting when a resource gap has opened in a project and deciding how to best utilize the project's resources to fill the resource gap, a processor monitors a degree of progress of one or more projects on a project dashboard. Responsive to detecting a project with a resource gap, a processor obtains a set of data regarding the project. A processor determines a productivity trajectory. Responsive to determining a team member will not meet the productivity trajectory, a processor determines one or more alternative combinations of team members who can meet the productivity trajectory. A processor measures a degree of amplification required to achieve the productivity trajectory by the one or more alternative combinations of team members. A processor identifies one or more resources available to complete a backlog of one or more tasks of the first project. A processor determines an optimum variation.

CLOSED LOOP VERIFICATION FOR EVENT GROUPING MECHANISMS (17809077)

Main Inventor

Pooja Aggarwal

Brief explanation

The patent application describes a method and system for using a causal dependence graph of events in a large enterprise system to identify the most frequently used corrective action for a set of actions required by the enterprise.

• The method utilizes a causal dependence graph to analyze the relationships between different events in the enterprise system.
• The system correlates a set of actions or workflows with a set of corrective actions.
• The method and system are designed to handle large sets of data related to actions performed by the enterprise system.
• The goal is to determine the most frequently utilized corrective action for a given set of actions.

Potential Applications

This technology can be applied in various industries and sectors where large enterprise systems are used, such as:

• Manufacturing: Identifying the most effective corrective actions for optimizing production processes.
• IT Operations: Determining the most common solutions for resolving system issues and errors.
• Customer Service: Finding the most frequently used corrective actions for addressing customer complaints or inquiries.
• Supply Chain Management: Optimizing logistics and inventory management by identifying the most effective corrective actions.

Problems Solved

The technology addresses the following problems:

• Difficulty in correlating a set of actions or workflows with the appropriate corrective actions in large enterprise systems.
• Inefficient and time-consuming manual analysis of data to determine the most frequently utilized corrective actions.
• Lack of a systematic approach to identify the most effective solutions for common issues or problems in enterprise systems.

Benefits

The use of this technology offers several benefits:

• Improved efficiency in identifying and implementing corrective actions in enterprise systems.
• Reduction in manual effort and time required for analyzing large sets of data.
• Enhanced decision-making by utilizing a systematic approach based on causal dependence analysis.
• Optimization of enterprise processes and workflows through the identification of frequently utilized corrective actions.

Abstract

A method and system is provided for utilizing a causal dependence graph of events in a large enterprise-related system to determine a most frequently utilized corrective action for a set of actions that the enterprise requires. Typically, with large sets of data related to actions that an enterprise system performs, it is non-trivial to correlate a set of actions (or workflows) with a set of corrective actions.

USING DATA ANALYTICS TO OPTIMIZE LOGISTICS WITHIN PRODUCT DISTRIBUTION NETWORK (17809305)

Main Inventor

Siddhartha Sood

Brief explanation

The abstract describes a logistics optimization system that uses data analytics to determine the best delivery plan for a product within a distribution network. The system analyzes order data, including order information, delivery location, and order history, to calculate the probability of product return and re-order. Based on this analysis, the system generates a recommended storage facility for a second delivery location in case the original order is canceled.

• The system optimizes logistics for delivering products within a distribution network.
• A computer receives order data, including order information and delivery location.
• The order data also includes the user's order history, including returns or cancelations.
• Data analytics is used to analyze the order data and determine the best delivery plan.
• The system calculates the probability of product return and re-order.
• Based on the analysis, a recommended storage facility is generated for a second delivery location in case of order cancelation.

Potential Applications

• E-commerce companies can use this system to optimize their logistics and improve delivery efficiency.
• Retailers can benefit from this system by reducing the impact of order cancelations and improving customer satisfaction.
• Logistics companies can use this system to streamline their operations and reduce costs.

Problems Solved

• Inefficient logistics planning within a product distribution network.
• Difficulty in predicting and managing order cancelations and returns.
• Lack of optimized storage facility recommendations for re-routed deliveries.

Benefits

• Improved delivery efficiency and customer satisfaction.
• Reduced costs and improved operational efficiency for logistics companies.
• Better management of order cancelations and returns.

Abstract

Logistics are optimized for delivering a product within a product distribution network. A computer receives order data, the order data includes order information and a delivery location of a product ordered by a user. The order data further includes an order history of the user including order returns or cancelations. Using data analytics, the order data is analyzed to determine a logistics plan for the delivery of the product. A probability of return of the product is calculated, and a probability of re-order of the product is calculated. As part of the logistics plan and based on the analyzing of the order data, a recommended storage facility is generated for a second delivery location resulting from a re-routing of the delivery of the product, in response to the order being canceled.

JOINT LEARNING OF TIME-SERIES MODELS LEVERAGING NATURAL LANGUAGE PROCESSING (17846359)

Main Inventor

Shikhar Kwatra

Brief explanation

The patent application describes methods, computer program products, and systems for maximizing renewals of purchase orders. It involves using a classifier machine learning model to identify relevant metrics for predicting whether customers will renew purchase orders. The method also includes applying a tone analyzer natural language processing (NLP) model to determine the current sentiments of customers and recommending which customers to pursue with additional resources based on their sentiments and risks of non-renewal.

• Utilizes a classifier machine learning model to identify relevant metrics for predicting purchase order renewals
• Predicts risks of non-renewal for purchase orders using the identified metrics
• Applies a tone analyzer natural language processing (NLP) model to determine the current sentiments of customers
• Recommends which customers to pursue with additional resources based on their sentiments and risks of non-renewal

Potential Applications

This technology can be applied in various industries and sectors where purchase orders and customer renewals are significant, such as:

• E-commerce platforms
• Subscription-based services
• Supply chain management
• Retail businesses

Problems Solved

This technology addresses the following problems:

• Difficulty in identifying relevant metrics for predicting purchase order renewals
• Lack of efficient methods to assess risks of non-renewal for purchase orders
• Inability to determine the sentiments of customers accurately
• Challenges in deciding which customers to focus on for maximizing renewals

Benefits

The use of this technology offers several benefits:

• Improved accuracy in predicting purchase order renewals
• Enhanced risk assessment for non-renewal of purchase orders
• Better understanding of customer sentiments through natural language processing
• Optimal allocation of resources towards customers with higher chances of renewal

Abstract

Disclosed are methods, computer program products, and systems for maximizing renewals of purchase orders. One embodiment of the method may comprise utilizing a classifier machine learning model to identify metrics that are most relevant to whether customers will renew purchase orders, predicting respective risks of non-renewal for the purchase orders using the identified metrics, applying a tone analyzer natural language processing (NLP) model to determine current sentiments for respective customers, and recommending which of the respective customers to pursue with additional resources based the respectively determined sentiments and risks of non-renewal.

PERSONALIZED ADAPTIVE EDUCATION COURSE ADAPTING TO NEURODIVERSITY CONSIDERATIONS IN ONLINE LEARNING (17808324)

Main Inventor

Martin G. Keen

Brief explanation

The patent application describes a method for customizing online educational course material to meet the specific needs of students with neurodivergent conditions. Here are the key points:

• The processor receives a request from a user to tailor the course material.
• It identifies the user's neurodivergent needs based on existing data.
• It observes the user in real-time as they complete the course to identify additional needs.
• It determines the user's preferences.
• It generates recommendations on how to customize the course material.
• It creates a personalized user interface for the user to access the tailored course material.

Potential applications of this technology:

• Online educational platforms can use this method to provide personalized learning experiences for students with neurodivergent conditions.
• Schools and universities can incorporate this method into their online courses to support students with diverse learning needs.
• Educational technology companies can develop software or tools based on this method to assist students with neurodivergent conditions.

Problems solved by this technology:

• Traditional online educational courses may not adequately address the unique learning needs of students with neurodivergent conditions.
• Students with neurodivergent conditions may struggle to succeed in online courses without appropriate learning supports.
• Educators may find it challenging to provide individualized assistance to students with diverse neurodivergent needs.

Benefits of this technology:

• Students with neurodivergent conditions can receive tailored learning supports to help them succeed in online courses.
• The method allows for real-time observation of students, enabling the identification of additional needs that may not be captured in existing data.
• Personalized user interfaces and customized course material can enhance the learning experience for students with neurodivergent conditions.

Abstract

In an approach for tailoring a set of course material of an online educational course to provide one or more learning supports necessary for a student with a neurodivergent-classified condition to succeed, a processor receives a request from a user to tailor the set of course material. A processor identifies one or more neurodivergent needs of the user based on a first set of data gathered from a database. A processor observes the user complete an unmodified version of the online educational course in real-time to identify one or more additional neurodivergent needs of the user. A processor determines one or more preferences of the user. A processor generates a set of recommendations on how to tailor the set of course material. A processor creates a personalized user interface based on the set of recommendations for the user to access a tailored set of course material.

INFORMATION EXTRACTION FROM DOCUMENTS CONTAINING HANDWRITTEN TEXT (17809321)

Main Inventor

Saurabh Goyal

Brief explanation

The present invention is a method, computer system, and computer program product for information extraction. It involves receiving a mixed-text document that contains both typed and handwritten text, including at least one key-value pair. The invention also includes receiving the location information of at least one key from the mixed-text document and detecting the handwritten text based on this location information.

• The invention is a system that can extract information from mixed-text documents containing both typed and handwritten text.
• It includes a handwriting detection model that can identify handwritten text within the document.
• The system can also identify the location of specific keys within the document.
• By combining the location information and handwriting detection, the system can accurately extract the handwritten text associated with each key-value pair.

Potential Applications

• This technology can be used in data entry tasks where information needs to be extracted from mixed-text documents.
• It can be applied in various industries such as finance, healthcare, and legal, where handwritten forms or documents are still commonly used.
• The system can automate the process of extracting information from mixed-text documents, saving time and reducing errors.

Problems Solved

• Manual extraction of information from mixed-text documents can be time-consuming and error-prone.
• Handwritten text is often difficult to interpret, leading to inaccuracies in data extraction.
• This technology solves these problems by automating the extraction process and accurately identifying handwritten text.

Benefits

• The system improves efficiency by automating the extraction of information from mixed-text documents.
• It reduces errors by accurately detecting and extracting handwritten text.
• The technology can be integrated into existing systems, enhancing their capabilities in handling mixed-text documents.

Abstract

A method, computer system, and a computer program product for information extraction is provided. The present invention may include receiving, by a handwriting detection model of an integrated system, a mixed-text document including a combination of typed text and handwritten text, where the received mixed-text document includes at least one key-value pair. The present invention may also include receiving, by the handwriting detection model of the integrated system, a first location information of at least one key from the at least one key-value pair in the received mixed-text document. The present invention may further include detecting, by the handwriting detection model of the integrated system, at least one handwritten text in the received mixed-text document based on the received first location information of the at least one key.

INTRUSION MOVEMENT PREDICTION (18461637)

Main Inventor

Doga Tav

Brief explanation

The patent application describes a method, computer system, and computer program for predicting intrusion movement in a monitored space using environmental sensor data. 

• The method involves receiving data from multiple sensors attached to an airflow component in the monitored space.
• The received data is used to generate a three-dimensional model of the monitored space.
• If a disturbance is detected in the three-dimensional model, a security action is triggered.

Potential Applications

This technology has potential applications in various fields, including:

• Security systems: It can be used to enhance intrusion detection and prevention in buildings, homes, and other monitored spaces.
• Industrial facilities: It can help identify and respond to unauthorized access or movement in sensitive areas.
• Retail stores: It can be used to prevent theft or shoplifting by detecting suspicious movements.

Problems Solved

This technology addresses the following problems:

• Limited accuracy: Traditional intrusion detection systems may have limitations in accurately detecting and predicting intrusions.
• False alarms: Existing systems may generate false alarms, leading to unnecessary security actions and wasted resources.
• Lack of real-time monitoring: Some systems may not provide real-time monitoring and response to potential intrusions.

Benefits

The use of this technology offers several benefits:

• Enhanced security: By predicting intrusion movement, potential threats can be identified and addressed proactively.
• Improved accuracy: The three-dimensional model generated using environmental sensor data provides a more accurate representation of the monitored space.
• Real-time response: The system can trigger security actions immediately upon detecting a disturbance, minimizing response time.

References

[Link to the patent application]

Abstract

According to one embodiment, a method, computer system, and computer program product for intrusion movement prediction is provided. The embodiment may include receiving environmental sensor data corresponding to a monitored space as captured by a plurality of sensors affixed to an airflow component. The embodiment may also include generating a three-dimensional model of the monitored space using the received environmental data. The method may further include, in response to determining a disturbance is present in the three-dimensional model, performing a security action.

RESPONSIVE LIVE MUSICAL SOUND GENERATION (17809173)

Main Inventor

Killian Levacher

Brief explanation

The patent application describes a system that generates real-time vocal audio for a song based on predetermined musical data. Here are the key points:

• The system receives predetermined musical data, which includes chords, lyrics, and rhythmic structures of a song.
• It also receives audio data of a band generating music for the song.
• The system then generates real-time vocal audio that is in rhythm with the audio data and in harmony with the chords.
• The vocal audio includes the lyrics and is of a predetermined voice.

Potential applications of this technology:

• Music production: This system can be used in music production to quickly generate vocal audio for songs, saving time and effort.
• Karaoke systems: The technology can be implemented in karaoke systems to provide real-time vocal audio for users singing along to songs.
• Virtual band performances: It can be used in virtual band performances, where the vocal audio is generated in real-time to match the music being played.

Problems solved by this technology:

• Time-consuming process: Generating vocal audio for a song can be a time-consuming task. This technology automates the process, making it more efficient.
• Ensuring harmony and rhythm: Creating vocal audio that is in harmony with the chords and in rhythm with the music can be challenging. This system solves this problem by using predetermined musical data.

Benefits of this technology:

• Efficiency: The system automates the process of generating vocal audio, saving time and effort for music producers.
• Accuracy: By using predetermined musical data, the system ensures that the vocal audio is in harmony with the chords and in rhythm with the music.
• Consistency: The predetermined voice used for the vocal audio ensures consistency across different songs and performances.

Abstract

Predetermined musical data for a song is received. The predetermined musical data includes chords and lyrics and rhythmic structures of the song. Audio data of a band generating music of the song is received. Generating real-time vocal audio that is in rhythm with the audio data and in harmony with the chords. The vocal audio includes the lyrics and is of a predetermined voice.

ARTIFICIAL INTELLIGENCE FACTSHEET GENERATION FOR SPEECH RECOGNITION (17809202)

Main Inventor

Shreya Khare

Brief explanation

The patent application describes a method, system, and computer program for automatically generating AI factsheets to customize speech to text (STT) models. Here are the key points:

• The method takes audio data containing human speech as input.
• It uses a first speech to text model to convert the audio data into text.
• The text data generated may contain errors, which are identified.
• AI factsheets are then generated to describe the model metadata of the first speech to text model.
• Based on the identified errors and the AI factsheets, a second speech to text model is generated specifically customized to the user.

Potential applications of this technology:

• Improving accuracy and customization of speech to text services.
• Enhancing transcription services for various industries, such as legal, medical, and media.
• Enabling better voice assistants and voice-controlled devices.
• Assisting in language learning and pronunciation improvement.

Problems solved by this technology:

• Reducing transcription errors in speech to text conversion.
• Providing customized speech to text models tailored to individual users.
• Streamlining the process of generating AI factsheets for model metadata.

Benefits of this technology:

• Improved accuracy and efficiency in converting speech to text.
• Customized models that better understand and transcribe individual users' speech patterns.
• Enhanced user experience with voice-controlled devices and applications.
• Time and cost savings in transcription services.

Abstract

A method, system, and computer program product for automated artificial intelligence (AI) factsheet generation for modeling and model customization in speech to text (STT) services. The method receives audio data for a user. The audio data contains human speech. Text data is generated, using a first speech to text model, to represent the human speech of the audio data. A set of transcription errors of the first speech to text model are identified. A set of AI factsheets are generated to describe model metadata for the first speech to text model. Based on the set of transcription errors and the set of AI factsheets, the method generates a second speech to text model customized to the user.

COMPUTER TECHNOLOGY FOR CONTROLLING DIGITAL TWIN SIMULATION IN VOICE ASSISTANT DEVICES (17809286)

Main Inventor

Shilpa Bhagwatprasad Mittal

Brief explanation

The abstract describes a computer technology that uses digital twin simulation to improve the precision and accuracy of query responses in a virtual assistant system. Here are the key points:

• The technology receives a query and a desired level of precision and/or accuracy associated with the query.
• It determines if the desired level exceeds a predetermined threshold.
• If the desired level exceeds the threshold, the system applies digital twin simulation to a digital twin data set to improve the precision and accuracy of the query response.
• The improved query response is then outputted to the querying party, for example, as sound data in a voice assistant system.
• If the desired level is below the threshold, the digital twin simulation is not performed, resulting in a quicker response and lower computational power demands on the virtual assistant system.
• The precision and accuracy of the query response can be calculated based on the context.

Potential applications of this technology:

• Voice assistant systems: This technology can be applied to improve the precision and accuracy of voice-based query responses, making voice assistants more reliable and effective.
• Chatbots: Chatbot systems can benefit from this technology by providing more accurate and precise responses to user queries, enhancing the user experience.
• Customer support: Virtual assistant systems used in customer support can utilize this technology to provide more accurate and helpful responses to customer queries, improving customer satisfaction.

Problems solved by this technology:

• Lack of precision and accuracy: By applying digital twin simulation, this technology addresses the issue of inaccurate or imprecise query responses, ensuring that users receive the most relevant and reliable information.
• Computational power demands: By selectively applying the simulation based on the desired level of precision and accuracy, this technology reduces the computational power required for query processing, making the system more efficient.

Benefits of this technology:

• Improved user experience: By enhancing the precision and accuracy of query responses, users can rely on the virtual assistant system to provide more helpful and relevant information.
• Faster response times: By skipping the simulation when the desired level is below the threshold, the system can provide quicker responses, reducing waiting times for users.
• Efficient resource utilization: The selective application of digital twin simulation helps optimize computational power usage, making the virtual assistant system more resource-efficient.

Abstract

Computer technology that receives a query and a desired level of precision and/or accuracy (herein, a “p/a value”) associated with the received query, determines whether the p/a value exceeds a predetermined threshold, if the p/a value exceeds the predetermined threshold, then the virtual assistant system applies digital twin simulation to a digital twin data set in determining the query response to improve precision and/or accuracy of the query response, and outputs the query response to the querying party (for example, output as sound data in a voice assistant system). If the p/a value is below the threshold value, then the digital twin simulation is not performed, which makes the response quicker and puts less computational power demands on the virtual assistant system. In some embodiments, the virtual assistant system calculates the p/a value based on context.

AI DRIVEN SMART PATIENT LABELING SYSTEM (17809416)

Main Inventor

Saigeetha Aswathnarayanan Jegannathan

Brief explanation

The abstract describes a method for automatically identifying updates in a Scientific Drug Label (SL) and incorporating them into a simplified Patient Drug Label (PL). The method involves converting complex medical language into patient-friendly language, identifying modified or new content in the SL, finding corresponding locations in the PL, and incorporating the updates into the PL.

• The method automatically identifies updates in a Scientific Drug Label (SL).
• The updates are incorporated into a simplified Patient Drug Label (PL).
• Complex medical language in the SL is converted into patient-friendly language.
• Modified, inserted, or deleted content in the SL is identified.
• Corresponding locations in the PL are found for the identified updates.
• The updates are incorporated into the mapped locations of the PL.

Potential Applications

• Pharmaceutical companies can use this method to automatically update patient drug labels with the latest information from scientific drug labels.
• Healthcare providers can benefit from having patient drug labels that are up-to-date and easy for patients to understand.

Problems Solved

• Manual updating of patient drug labels can be time-consuming and prone to errors.
• Complex medical language in scientific drug labels can be difficult for patients to understand.
• Ensuring that patient drug labels accurately reflect the latest information from scientific drug labels can be challenging.

Benefits

• Automation of the updating process saves time and reduces the risk of errors.
• Patient drug labels are simplified and easier for patients to comprehend.
• Patient drug labels are kept up-to-date with the latest information from scientific drug labels.

Abstract

In an approach for automatically identifying one or more updates in a Scientific Drug Label (SL) relevant to a patient and incorporating the one or more updates into a Patient Drug Label (PL), a processor receives a pair of documents, wherein the pair of documents include the SL and the PL. A processor converts a complex medical language of the SL into a simplified patient friendly language. A processor identifies one or more words, one or more phrases, or one or more sentences that have been modified, inserted, or deleted. A processor searches for a location in the PL that closely maps to the one or more words, the one or more phrases, or the one or more sentences to the SL. A processor incorporates the one or more words, the one or more phrases, or the one or more sentences in a mapped location of the PL.

INTERCONNECT WITH TWO-DIMENSIONAL FREE ZERO LINE END ENCLOSURE (17808316)

Main Inventor

Ruilong Xie

Brief explanation

The patent application describes a technology for providing interconnects with a specific type of enclosure. The interconnects consist of two metal lines connected by a via. 

• The first metal line has a zero line extension at one end in a certain dimension.
• The second metal line has a zero line extension at one end in a perpendicular dimension to the first dimension.

Potential applications of this technology:

• Integrated circuits and semiconductor devices
• Electronics manufacturing and assembly
• Communication systems and networks

Problems solved by this technology:

• Ensures proper connection and signal transmission between metal lines
• Reduces signal interference and noise
• Improves overall performance and reliability of interconnects

Benefits of this technology:

• Enhanced signal integrity and quality
• Increased efficiency and speed of data transmission
• Improved reliability and durability of interconnects

Abstract

Embodiments of the invention include providing interconnects with two-dimensional free zero line end enclosure. A first metal line is formed. A second metal line is connected by a via to the first metal line, the first metal line having a first end with a zero line extension in relation to the via in a first dimension, the second metal line having another first end with a zero line extension in relation to the via in a second dimension perpendicular to the first dimension.

ENHANCED POWER AND SIGNAL FOR STACKED-FETS (17846423)

Main Inventor

Albert M. Young

Brief explanation

==Abstract==

A semiconductor structure has been developed to address the shorting risks associated with using long bars in advanced logic applications. This structure, which utilizes stacked field effect transistor technology with unequal device footprints, provides a reliable power rail that can mitigate or eliminate these risks.

Patent/Innovation Explanation

• The semiconductor structure is designed for advanced logic applications.
• It utilizes stacked field effect transistor technology.
• The structure includes unequal device footprints to address shorting risks.
• A reliable power rail is provided to mitigate or eliminate these risks.

Potential Applications

This technology has potential applications in various fields, including:

• Advanced logic applications
• Semiconductor manufacturing
• Electronics industry

Problems Solved

The semiconductor structure addresses the following problems:

• Shorting risks associated with using long bars in advanced logic applications
• Reliability issues in power rails

Benefits

The benefits of this technology include:

• Mitigation or elimination of shorting risks
• Improved reliability in power rails
• Enhanced performance in advanced logic applications

Abstract

A semiconductor structure including a reliable power rail in stacked field effect transistor technology with unequal device footprints is provided that mitigates, and in some cases even eliminates, shorting risks that are typically associated using long bars in advanced logic applications.

SEMICONDUCTOR DEVICE WITH POWER VIA (17808186)

Main Inventor

Ruilong Xie

Brief explanation

The abstract describes a semiconductor device that includes a field effect transistor (FET) with source/drain (S/D) epitaxial regions, a gate cut region, a dielectric liner and core, a backside power rail (BPR) and distribution network (BSPDN), metal contacts, and a via to backside power rail (VBPR) contact. The dielectric liner separates the power via from the first S/D epitaxial region.

• The semiconductor device includes a field effect transistor (FET) with first and second source/drain (S/D) epitaxial regions.
• A gate cut region is present at cell boundaries between the first and second S/D epitaxial regions.
• A dielectric liner and core are formed in the gate cut region.
• A backside power rail (BPR) and a backside power distribution network (BSPDN) are included in the device.
• A power via passes through the dielectric core and connects to the BPR and BSPDN.
• First metal contacts are formed in contact with the first and second S/D epitaxial regions.
• A via to backside power rail (VBPR) contact is present.
• The dielectric liner separates the power via from the first S/D epitaxial region.

Potential Applications

• This semiconductor device can be used in various electronic devices that require efficient power distribution and management.

Problems Solved

• The device solves the problem of power distribution and management in a semiconductor device by providing a backside power rail and distribution network.

Benefits

• The device allows for efficient power distribution and management, improving the overall performance and reliability of electronic devices.

Abstract

A semiconductor device is provided. The semiconductor device includes a field effect transistor (FET) including first and second source/drain (S/D) epitaxial regions. The semiconductor device also includes a gate cut region at cell boundaries between the first and second S/D epitaxial regions, a dielectric liner and a dielectric core formed in the gate cut region, and a backside power rail (BPR) and a backside power distribution network (BSPDN). The semiconductor device also includes a power via passing through the dielectric core and connecting to the BPR and BSPDN, first metal contacts formed in contact with the first and second S/D epitaxial regions, and a via to backside power rail (VBPR) contact. The dielectric liner separates the power via from the first S/D epitaxial region.

POWER PLANES AND PASS-THROUGH VIAS (17808116)

Main Inventor

Nicholas Anthony Lanzillo

Brief explanation

The semiconductor device described in the patent application consists of multiple layers of metal and dielectric materials. 

• The first metal layer contains signal tracks and power rails embedded in a dielectric layer.
• The second metal layer also contains signal tracks and power rails embedded in a dielectric layer.
• A metal plane is positioned between the first and second metal layers.
• A third dielectric layer separates the first metal layer from the metal plane.
• A fourth dielectric layer separates the second metal layer from the metal plane.

Potential applications of this technology:

• Integrated circuits
• Microprocessors
• Memory devices
• Power management systems

Problems solved by this technology:

• Efficient signal transmission and power distribution within a semiconductor device
• Reduction of signal interference and noise
• Improved performance and reliability of electronic systems

Benefits of this technology:

• Enhanced functionality and performance of semiconductor devices
• Higher data transfer rates and faster processing speeds
• Improved power efficiency and reduced energy consumption
• Increased reliability and durability of electronic systems.

Abstract

The semiconductor device includes a first metal layer, a second metal layer, a metal plane, a third dielectric layer and a fourth dielectric layer. The first metal layer comprises a first dielectric layer with a first plurality of signal track and a first plurality of power rails. The second metal layer comprises a second dielectric layer with a second plurality of signal tracks and a second plurality of power rails. The metal plane is between the first metal layer and the second metal layer. The third dielectric layer is between the first metal layer and the metal plane. The fourth dielectric layer is between the second metal layer and the metal plane.

CONTACTS FOR STACKED FIELD EFFECT TRANSISTOR (17808568)

Main Inventor

Ruilong Xie

Brief explanation

Abstract:

A semiconductor device is described in this patent application. The device consists of a first field effect transistor (FET) and a second FET stacked on top of the first one. It also includes a backside contact (BSCA) that is connected to a backside power rail (BSPR). Additionally, there is a via to backside power rail (VBPR) that lands over the BSCA.

Patent/Innovation Explanation:

• The semiconductor device includes two stacked field effect transistors (FETs).
• A backside contact (BSCA) is connected to a backside power rail (BSPR).
• A via to backside power rail (VBPR) is present, landing over the BSCA.

Potential Applications:

• This semiconductor device can be used in various electronic devices such as smartphones, computers, and other consumer electronics.
• It can be utilized in power management circuits, amplifiers, and other integrated circuits.

Problems Solved:

• The stacked FETs provide improved performance and functionality in electronic devices.
• The backside contact and power rail connection enhance power distribution and efficiency.
• The via to backside power rail allows for efficient power delivery to the device.

Benefits:

• The stacked FETs offer increased circuit density and performance.
• The backside contact and power rail connection improve power distribution and reduce power losses.
• The via to backside power rail enables efficient power delivery, enhancing overall device efficiency.

Abstract

A semiconductor device is provided. The semiconductor device includes a first field effect transistor (FET); a second FET stacked over the bottom FET; a backside contact (BSCA) connected to a backside power rail (BSPR); and a via to backside power rail (VBPR), the VBPR landing over the BSCA.

STACKED FIELD EFFECT TRANSISTOR CELL WITH CROSS-COUPLING (17809076)

Main Inventor

Carl Radens

Brief explanation

The abstract describes a patent application for a complementary metal oxide semiconductor (CMOS) device with a hybrid cross-couple contact. The contact includes a frontside contact to the gate of the CMOS device, a source contact to the source of the CMOS device, and a drain contact to the drain of the CMOS device. The frontside contact is located on the frontside of the device, while the source and drain contacts are located on the backside.

• The patent application is for a CMOS device with a hybrid cross-couple contact.
• The hybrid cross-couple contact includes a frontside contact, a source contact, and a drain contact.
• The frontside contact is connected to the gate of the CMOS device and is located on the frontside of the device.
• The source contact is connected to the source of the CMOS device and is located on the backside of the device.
• The drain contact is connected to the drain of the CMOS device and is also located on the backside of the device.

Potential Applications

• This technology can be used in the manufacturing of CMOS devices.
• It can be applied in various electronic devices that utilize CMOS technology, such as smartphones, computers, and other consumer electronics.

Problems Solved

• The hybrid cross-couple contact solves the problem of connecting the gate, source, and drain of a CMOS device efficiently.
• It provides a simplified and compact design for the CMOS device.

Benefits

• The hybrid cross-couple contact allows for improved performance and functionality of CMOS devices.
• It enables better integration and miniaturization of electronic devices.
• The frontside contact and backside contacts provide a more efficient and effective connection for the CMOS device.

Abstract

Embodiments are disclosed for a complementary metal oxide semiconductor (CMOS) device. The CMOS device includes a hybrid cross-couple contact. The hybrid cross-couple contact includes a frontside contact to a gate of the CMOS device. The frontside contact is disposed on a frontside of the CMOS device. The hybrid cross-couple contact includes a source contact to a source of the CMOS device. The source contact is disposed on a backside of the CMOS device. The hybrid cross-couple contact includes a drain contact to a drain of the CMOS device. The drain contact is disposed on a backside of the CMOS device.

ELECTROMAGNETIC WAVEGUIDING THROUGH LIQUID COOLING CONDUIT (17808221)

Main Inventor

Anil Yuksel

Brief explanation

The abstract describes a multi-chip package structure that includes two IC chips and a fluid conduit. The structure is designed to remove heat generated by the IC chips. The fluid conduit is thermally coupled to both IC chips and allows for the transmission of electromagnetic signals between them.

• The multi-chip package structure includes two IC chips and a fluid conduit.
• The fluid conduit is thermally coupled to both IC chips.
• The structure is designed to remove heat generated by the IC chips.
• A first monopole feed is connected between the first IC chip and one end of the fluid conduit.
• A second monopole feed is connected between the second IC chip and the other end of the fluid conduit.
• The first monopole feed transmits electromagnetic signals through the fluid conduit to the second monopole feed, and vice versa.

Potential Applications

• This technology can be applied in various electronic devices that require efficient heat dissipation, such as computers, servers, and mobile devices.
• It can be used in high-performance computing systems where multiple IC chips generate significant heat.
• The multi-chip package structure can be utilized in telecommunications equipment to improve thermal management.

Problems Solved

• The multi-chip package structure solves the problem of heat accumulation in IC chips by providing an efficient heat removal mechanism.
• It addresses the challenge of maintaining optimal operating temperatures in electronic devices with multiple IC chips.
• The technology solves the issue of electromagnetic signal transmission between IC chips in a thermally coupled system.

Benefits

• The structure allows for effective heat dissipation, preventing overheating and potential damage to IC chips.
• It enables efficient thermal management in electronic devices, improving their overall performance and reliability.
• The transmission of electromagnetic signals through the fluid conduit enhances communication and data transfer between IC chips.

Abstract

A multi-chip package structure is provided. The multi-chip package structure includes a first IC chip and a second IC chip, and a fluid conduit thermally coupled to the first IC chip and the second IC chip. The multi-chip package structure is configured to remove heat generated by at least one of the first IC chip and the second IC chip. The fluid conduit has a first end and a second end opposite to the first end. The multi-chip package structure also includes a first monopole feed connected between the first IC chip and the first end of the fluid conduit, and a second monopole feed connected between the second IC chip and the second end of the fluid conduit. The first monopole feed is configured to transmit an electromagnetic signal through the fluid conduit toward the second monopole feed and vice versa.

SINGLE STACK DUAL CHANNEL GATE-ALL-AROUND NANOSHEET WITH STRAINED PFET AND BOTTOM DIELECTRIC ISOLATION NFET (17808360)

Main Inventor

Julien Frougier

Brief explanation

The abstract describes a patent application for a new type of nanosheet transistor with two channels, one made of silicon germanium (SiGe) and the other made of silicon (Si). The SiGe channel is strained to enhance its performance, while the Si channel is isolated from the bottom dielectric layer. 

• The invention is a single stack dual channel gate-all-around nanosheet transistor.
• It includes a strained PFET (p-type field-effect transistor) with a silicon germanium channel.
• It also includes a bottom dielectric isolation NFET (n-type field-effect transistor) with a silicon channel.
• The PFET and NFET are positioned laterally to each other.
• The silicon channel and the silicon germanium channel are staggered in a vertical direction.

Potential Applications

This technology has potential applications in various electronic devices and systems, including:

• Integrated circuits (ICs) for computers, smartphones, and other electronic devices.
• High-performance processors and memory chips.
• Power-efficient devices for Internet of Things (IoT) applications.
• Advanced sensors and imaging devices.

Problems Solved

The patent addresses several challenges in transistor design and fabrication, including:

• Enhancing the performance of p-type transistors by using strained silicon germanium channels.
• Isolating the n-type transistors from the bottom dielectric layer to prevent leakage and improve efficiency.
• Optimizing the layout and arrangement of the dual-channel transistors to maximize performance and minimize interference.

Benefits

The proposed technology offers several benefits over traditional transistor designs:

• Improved performance and efficiency due to the strained silicon germanium channel in the p-type transistor.
• Enhanced isolation and reduced leakage in the n-type transistor through bottom dielectric isolation.
• Compatibility with existing fabrication processes, allowing for easier integration into current manufacturing workflows.
• Potential for higher transistor density and smaller chip sizes, leading to more compact and powerful electronic devices.

Abstract

Embodiments of the invention include a single stack dual channel gate-all-around nanosheet with strained PFET and bottom dielectric isolation NFET. A PFET comprising at least one silicon germanium channel is formed. An NFET comprising at least one silicon channel is formed, the PFET being positioned laterally to the NFET, the at least one silicon channel and the at least one silicon germanium channel being staggered in a vertical direction.

DUMMY FIN CONTACT IN VERTICALLY STACKED TRANSISTORS (17847765)

Main Inventor

Joshua M. Rubin

Brief explanation

The abstract describes a patent application for a vertically stacked transistor structure within a wafer. The structure includes multiple transistor components, such as a first bottom transistor with a channel, gate, source, and drain, and a first contact connected to the first bottom transistor. It also includes a set of stacked transistors, with a second top transistor on top of a second bottom transistor.

• The patent application describes a vertically stacked transistor structure within a wafer.
• The structure includes multiple transistor components, such as a first bottom transistor with a channel, gate, source, and drain.
• A first contact is connected to the first bottom transistor.
• The structure also includes a set of stacked transistors, with a second top transistor on top of a second bottom transistor.

Potential Applications

• This technology could be used in the manufacturing of integrated circuits and microprocessors.
• It may enable the creation of more compact and efficient electronic devices.
• The vertically stacked transistor structure could be utilized in various industries, including telecommunications, computing, and consumer electronics.

Problems Solved

• The vertically stacked transistor structure allows for increased transistor density within a wafer, addressing the need for smaller and more powerful electronic devices.
• It provides a solution for improving the performance and efficiency of integrated circuits and microprocessors.
• The patent application offers a method for forming the vertically stacked transistor structure, solving the challenge of manufacturing such complex structures.

Benefits

• The vertically stacked transistor structure allows for more transistors to be packed into a smaller space, leading to higher performance and functionality in electronic devices.
• It enables the creation of more compact and energy-efficient integrated circuits and microprocessors.
• The method described in the patent application provides a reliable and scalable approach to forming the vertically stacked transistor structure.

Abstract

A plurality of transistor components, a system, and a method of forming a vertically stacked transistor structure within a wafer. The plurality of transistor components may include a first bottom transistor, where the first bottom transistor includes a channel, a gate, a source, and a drain. The plurality of transistor components may also include a first contact on top of the first bottom transistor, where the first contact is proximately connected to the first bottom transistor. The plurality of transistor components may also include a first set of stacked transistors, where the first set of stacked transistors includes a second top transistor on top of a second bottom transistor.

Ferroelectric Film with Buffer Layers for Improved Reliability of Metal-Insulator-Metal Capacitor (17851290)

Main Inventor

Kisik Choi

Brief explanation

The patent application describes a new design for metal-insulator-metal capacitors that have increased reliability. The capacitors include multiple dielectric layers between two electrodes, with a ferroelectric film in the middle. The ferroelectric film is made up of a combination of two dielectric materials with different properties. The first and second dielectric materials can be HfO and/or ZrO in a crystalline phase, either combined in a single layer or present in separate layers. The capacitors can be stacked on top of each other to create a capacitor device. The method of forming these capacitors is also described.

• Metal-insulator-metal capacitors with increased reliability
• Multiple dielectric layers between electrodes
• Ferroelectric film made up of a combination of two dielectric materials
• Dielectric materials can be HfO and/or ZrO in a crystalline phase
• Capacitors can be stacked to create a capacitor device
• Method of forming the capacitors is provided

Potential Applications

• Electronics industry
• Energy storage systems
• Integrated circuits

Problems Solved

• Increased reliability of metal-insulator-metal capacitors
• Improved performance of capacitors in electronic devices

Benefits

• Enhanced reliability of capacitors
• Improved performance of electronic devices
• Increased energy storage capacity

Abstract

Metal-insulator-metal capacitor designs with increased reliability are provided. In one aspect, a capacitor includes: first and second electrodes; and multiple dielectric layers present in between the first and second electrodes, including a first buffer layer disposed on the first electrode, a ferroelectric film disposed on the first buffer layer, and a second buffer layer disposed on the ferroelectric film, where the ferroelectric film includes a combination of at least a first dielectric material and a second dielectric material having a higher κ value than either the first or second buffer layers. The first and second dielectric materials can each include HfOand/or ZrO, in a crystalline phase, which can be combined in a common layer, or present in different layers. A capacitor device having the present capacitors stacked one on top of another is also provided, as is a method of forming the present capacitors.

CPP-AGNOSTIC SOURCE-DRAIN CONTACT FORMATION FOR GATE-ALL-AROUND DEVICES WITH DIELECTRIC ISOLATION (17850475)

Main Inventor

Julien Frougier

Brief explanation

The patent application describes a semiconductor structure that includes a source/drain (S/D) epitaxial growth formed over a bottom dielectric isolation region. It also includes at least one first semiconductor layer within the S/D region and at least one second semiconductor layer partially within a gate region. The second semiconductor layer extends into a spacer region to connect with the S/D epitaxial growth. The structure has different gate-to-gate spaces in different regions.

• The semiconductor structure includes S/D epitaxial growth over a dielectric isolation region.
• It has first and second semiconductor layers within the S/D and gate regions.
• The second semiconductor layer extends into a spacer region to connect with the S/D epitaxial growth.
• The structure has different gate-to-gate spaces in different regions.

Potential Applications

This technology can be applied in various semiconductor devices and integrated circuits, including but not limited to:

• Microprocessors
• Memory devices
• Power amplifiers
• Sensors

Problems Solved

The semiconductor structure addresses the following problems:

• Ensuring proper connection between the second semiconductor layer and the S/D epitaxial growth.
• Managing gate-to-gate spacing in different regions to optimize device performance.
• Providing a reliable and efficient structure for semiconductor devices.

Benefits

The benefits of this technology include:

• Improved connectivity between different semiconductor layers.
• Enhanced device performance through optimized gate-to-gate spacing.
• Reliable and efficient structure for semiconductor devices.
• Potential for improved functionality and integration in various applications.

Abstract

A semiconductor structure is presented including source/drain (S/D) epitaxial growth formed over a bottom dielectric isolation region, at least one first semiconductor layer disposed within the S/D epitaxial growth in a S/D region and at least one second semiconductor layer disposed partially within a gate region. The at least one second semiconductor layer extends from the gate region into a spacer region to enable a connection to the S/D epitaxial growth. The semiconductor structure further includes a first region with adjacent devices exhibiting a first Contacted gate Poly Pitch (CPP) defining a first gate-to-gate space and a second region with adjacent devices exhibiting a second CPP defining a second gate-to-gate space, where adjacent devices exhibiting the first CPP have a smaller gate-to-gate canyon than the adjacent devices exhibiting the second CPP such that the second gate-to-gate space is greater than the first gate-to-gate space.

SUBTRACTIVE SOURCE DRAIN CONTACT FOR STACKED DEVICES (17808124)

Main Inventor

Heng Wu

Brief explanation

The abstract describes a field effect transistor (FET) stack and a method for forming a contact to the lower source drain of the lower FET. The contact is adjacent to a vertical side surface of the lower source drain and has a reverse taper metal stud profile. A silicide layer is formed between the contact and the lower source drain.

• The FET stack includes a lower FET and an upper FET.
• A contact is formed to the lower source drain of the lower FET.
• The contact is adjacent to a vertical side surface of the lower source drain.
• A silicide layer is formed between the contact and the lower source drain.
• The overlap region between the silicide and the contact is smaller than the overlap region between the silicide and the first source drain.

Potential applications of this technology:

• Integrated circuits and semiconductor devices
• Electronics industry
• Computing and communication systems

Problems solved by this technology:

• Improved contact formation to the lower source drain of a lower FET
• Enhanced performance and efficiency of FET stacks

Benefits of this technology:

• Improved electrical conductivity and reliability of the contact
• Increased performance and efficiency of FET stacks
• Enables better integration and miniaturization of electronic devices.

Abstract

A field effect transistor (“FET”) stack, including a lower FET, and an upper FET, a first contact to a lower source drain of the lower FET, a first silicide between the first contact and the lower source drain, the first contact is adjacent to a vertical side surface of the lower source drain, a first overlap region between the first silicide and the first contact is less than a second overlap region between the first silicide and the first source drain. The first contact has a reverse tapper metal stud profile. Forming a first contact to a lower source drain of a lower FET of an FET stack, forming a first silicide between the first contact and the lower source drain, the first contact is adjacent to a vertical side surface of the lower source drain.

COMMON SELF ALIGNED GATE CONTACT FOR STACKED TRANSISTOR STRUCTURES (17808566)

Main Inventor

Su Chen Fan

Brief explanation

==Abstract==

A stacked semiconductor structure is described in this patent application. The structure consists of a top transistor placed above a bottom transistor, with a single gate contact that is in electrical contact with both the top gate conductor of the top transistor and the bottom gate conductor of the bottom transistor.

Patent/Innovation

• Stacked semiconductor structure with a top transistor and a bottom transistor.
• Single gate contact in electrical contact with both transistors.
• Allows for improved integration and compactness of semiconductor devices.

Potential Applications

This technology has potential applications in various fields, including:

• Integrated circuits
• Microprocessors
• Memory devices
• Power electronics
• Communication devices

Problems Solved

The stacked semiconductor structure solves several problems, such as:

• Limited space for integrating multiple transistors in a compact manner.
• Difficulty in achieving efficient electrical contact between stacked transistors.
• Complex fabrication processes for stacked semiconductor devices.

Benefits

The benefits of this technology include:

• Increased integration density, enabling more functionality in a smaller space.
• Improved electrical contact between stacked transistors, enhancing overall device performance.
• Simplified fabrication processes, reducing manufacturing complexity and cost.

Abstract

A stacked semiconductor structure including a top transistor stacked above a bottom transistor, and a single gate contact in electrical contact with a top gate conductor of the top transistor and a bottom gate conductor of the bottom transistor.

FORMING A FORKSHEET NANODEVICE (17849639)

Main Inventor

Ruilong Xie

Brief explanation

The patent application describes a semiconductor structure that includes a common substrate and two different types of complementary metal oxide semiconductor (CMOS) devices.

• The first CMOS device, called a forksheet device, is located on the common substrate and consists of an n-doped Field Effect Transistor (nFET) and a p-doped Field Effect Transistor (pFET). The effective width ratio (β) between the nFET and the pFET is defined.

• The second CMOS device, called a gate-all-around (GAA) nanosheet CMOS device, is adjacent to the forksheet device on the common substrate. It is a different type of CMOS device and has a different β value between its nFET and pFET.

• The second β value is at least 5 percent different from the first β value.

• The innovation lies in the combination of these two different CMOS devices with different β values on a common substrate.

Potential applications of this technology:

• This semiconductor structure can be used in various electronic devices, such as smartphones, tablets, computers, and other integrated circuits.

• It can improve the performance and efficiency of these electronic devices by utilizing the different characteristics of the two CMOS devices.

Problems solved by this technology:

• By combining different CMOS devices with different β values, the semiconductor structure can provide enhanced functionality and performance compared to traditional CMOS structures.

• It allows for better optimization of the semiconductor structure for specific applications and requirements.

Benefits of this technology:

• Improved performance and efficiency of electronic devices.

• Enhanced functionality and versatility of the semiconductor structure.

• Better optimization and customization for specific applications.

Abstract

A semiconductor structure includes a common substrate; a first forksheet complementary metal oxide semiconductor (CMOS) device that is located on the common substrate and that has an nFET (n-doped Field Effect Transistor) and a pFET (p-doped Field Effect Transistor) and has a first β (effective width ratio) between the nFET and the pFET; and a second forksheet device that is adjacent to the first forksheet device on the common substrate and that has a second β between a second nFET and a second pFET. The second β is different than the first β by at least 5 percent. Another semiconductor structure includes a common substrate; a forksheet complementary metal oxide semiconductor (CMOS) device that is located on the common substrate; and a gate-all-around (GAA) nanosheet CMOS device that is located on the common substrate and is adjacent to the forksheet device.

MAGNETIC STRUCTURES FOR RESONANT MANIPULATION OF SPIN (17808972)

Main Inventor

Michele Aldeghi

Brief explanation

The patent application describes a qubit system for quantum computing that utilizes a semiconductor structure, an array of plunger gates, and an array of magnetic structures. Here are the key points:

• The qubit system consists of a semiconductor structure, an array of plunger gates, and an array of magnetic structures.
• The plunger gates form a linear one-dimensional array of quantum dots (QDs) in the semiconductor structure.
• The magnetic structures generate stray fields in the same plane as the array of QDs.
• The QDs are positioned between the poles of individual magnetic structures.
• An external field is applied parallel to the linear array of QDs.
• The external field is adjusted to allow the magnetization of the magnetic structure to create a stray field, resulting in different total magnetic fields at different qubit locations.

Potential Applications:

• Quantum computing: The qubit system can be used for quantum computing applications, taking advantage of the unique properties of quantum dots and magnetic structures.

Problems Solved:

• Qubit stability: The use of magnetic structures and external fields helps stabilize the qubits, reducing the impact of external noise and fluctuations.

Benefits:

• Improved qubit performance: The qubit system's design and configuration enhance the stability and reliability of qubits, leading to improved performance in quantum computing applications.
• Scalability: The linear one-dimensional array of quantum dots allows for scalability, enabling the construction of larger and more powerful quantum computing systems.

Abstract

A qubit system for quantum computing includes a semiconductor structure, an array of plunger gates, and an array of magnetic structures. The array of gates is above the semiconductor structure forming a linear one-dimensional (1D) array of quantum dots (QDs) in the semiconductor structure. The array of magnetic structures generates stray fields in the same plane as the array of QDs. The QDs in the array are positioned between poles of individual magnetic structures in the array of magnetic structures. An external field is applied in a direction that is parallel to the linear 1D array of QDs. The external field is adjusted to allow the magnetization of the magnetic structure to create a stray field that leads to different total magnetic fields at different qubit locations.

OPTIMAL PROFILE SELECTION FOR FHE BASED ANALYTICAL MODELS (17809036)

Main Inventor

OMRI SOCEANU

Brief explanation

The patent application describes a method and system for evaluating and selecting the best packing solution for data that is processed through a fully homomorphic encryption (FHE) simulation. The goal is to find the most efficient way to pack the data for optimal performance and cost-effectiveness.

• The method involves running a simulation of the FHE system with different packing configurations.
• A user-selected model architecture can be provided to simulate various potential configurations.
• The cost of each simulated configuration is considered when determining the optimal packing solution.
• The system takes into account factors such as data size, computational requirements, and cost to evaluate the performance of different packing solutions.
• The method aims to find the most efficient packing solution that minimizes computational resources and cost while maintaining data integrity and security.

Potential applications of this technology:

• Secure cloud computing: The method can be applied to optimize the packing of data in secure cloud environments, ensuring efficient processing while maintaining data privacy.
• Data analytics: By finding the optimal packing solution, the method can enhance the performance of data analytics processes that involve fully homomorphic encrypted data.
• Machine learning: The technology can be used to improve the efficiency and cost-effectiveness of machine learning algorithms that operate on encrypted data.

Problems solved by this technology:

• Efficient packing: The method solves the problem of finding the most efficient packing solution for data processed through fully homomorphic encryption, optimizing performance and cost.
• Model architecture selection: By allowing users to provide a model architecture, the method addresses the challenge of simulating and evaluating different potential configurations.
• Cost optimization: Considering the cost of each simulated configuration helps solve the problem of minimizing computational resources and cost while maintaining data security.

Benefits of this technology:

• Enhanced performance: By finding the optimal packing solution, the method improves the performance of fully homomorphic encryption simulations, leading to faster processing times.
• Cost-effectiveness: Considering the cost of each configuration allows for cost optimization, resulting in more efficient resource utilization and reduced expenses.
• Customizability: Allowing users to provide a model architecture enables customization and flexibility in simulating and evaluating different packing configurations.

Abstract

A method and system for evaluating and selecting an optimal packing solution (or solutions) for data that is run through a fully homomorphic encryption (FHE) simulation. In some instances, a user selected model architecture is provided in order to start simulating multiple potential configurations. Additionally, the cost of each simulated configuration is taken into account when determining an optimal packing solution.

HIGHLY COLLABORATIVE DECEPTIVE NETWORK ALLIANCE (17808554)

Main Inventor

Doga Tav

Brief explanation

The abstract of this patent application describes a method, computer system, and computer program product for preventing intrusions on a network. The invention involves creating a sandbox environment specifically designed for an attacker detected on the network and then moving the attacker into this environment.

• The invention generates a sandbox environment in response to detecting an attacker on the network.
• The sandbox environment is customized with dynamically generated data tailored to the target of the attacker.
• The attacker is then isolated and moved into the sandbox environment.

Potential Applications

This technology can be applied in various areas where network security is crucial, such as:

• Corporate networks: Protecting sensitive data and preventing unauthorized access.
• Government networks: Safeguarding classified information and defending against cyber threats.
• Financial institutions: Securing financial transactions and preventing fraud.
• Healthcare systems: Protecting patient data and ensuring privacy.
• E-commerce platforms: Preventing unauthorized access to customer information and securing online transactions.

Problems Solved

The technology addresses the following problems in network security:

• Intrusion prevention: By isolating attackers in a sandbox environment, the system prevents them from accessing critical network resources.
• Targeted defense: The dynamically generated data in the sandbox environment is tailored to the attacker's target, making it more difficult for them to exploit vulnerabilities.
• Early detection: The system detects attackers on the network, allowing for proactive measures to be taken before any significant damage occurs.
• Risk mitigation: By moving attackers to a controlled environment, the risk of them causing harm to the network or stealing sensitive information is minimized.

Benefits

The use of this technology offers several benefits:

• Enhanced network security: By isolating attackers in a sandbox environment, the system provides an additional layer of protection against intrusions.
• Customized defense: The dynamically generated data in the sandbox environment makes it more challenging for attackers to achieve their objectives.
• Proactive approach: The system detects and responds to attackers in real-time, allowing for immediate action to prevent further damage.
• Reduced impact: By containing attackers in a controlled environment, the potential damage they can cause to the network and its resources is limited.

Abstract

According to one embodiment, a method, computer system, and computer program product for preventing intrusions on a network is provided. The present invention may include generating a sandbox environment responsive to detecting an attacker in the network, wherein the sandbox environment comprises dynamically generated data tailored to the target of the attacker; and moving the attacker to the sandbox environment.

DYNAMICALLY FEDERATED DATA BREACH DETECTION (17809021)

Main Inventor

Divyesh Jadav

Brief explanation

The abstract describes a system where a trained machine learning model and feature information are distributed from a server to multiple client devices. Each client device trains its own machine learning model using local data and constructs an unsupervised model using the feature information. The system then determines when there is a significant difference in performance between the supervised and unsupervised models and proposes changes to the feature information. If the proposed change improves the performance of a client device, it is communicated to a sample set of devices. If the majority of the sampled devices show improved performance, the change is communicated back to the server.

• A trained machine learning model and feature information are distributed from a server to multiple client devices.
• Each client device trains its own machine learning model using local data.
• Each client device constructs an unsupervised model using the feature information.
• The system determines when there is a significant performance difference between the supervised and unsupervised models.
• Proposed changes to the feature information are identified.
• The proposed change is deployed on one client device.
• If the proposed change improves the performance of that client device, it is communicated to a sample set of devices.
• If the majority of the sampled devices show improved performance, the change is communicated back to the server.

Potential Applications

This technology can be applied in various domains where distributed machine learning models are used, such as:

• Fraud detection systems
• Anomaly detection in network security
• Predictive maintenance in industrial settings
• Personalized recommendation systems

Problems Solved

This technology addresses the following problems:

• Centralized training of machine learning models may not be feasible or efficient for large-scale deployments.
• Local data on client devices may have unique characteristics that can improve the performance of the models.
• The detection performance of supervised and unsupervised models may differ significantly, requiring adjustments to the feature information.

Benefits

The benefits of this technology include:

• Improved detection performance by leveraging local data and unsupervised models.
• Efficient distribution of trained models and feature information to client devices.
• Adaptive system that can identify and communicate changes that improve performance.
• Reduced reliance on centralized server for model training and updates.

Abstract

A processor distributes, from a server, a trained supervised machine learning (ML) model and supervised and unsupervised feature information to a plurality of client devices; at each client device, trains the supervised ML model using local data to generate a local supervised ML model, constructs a local unsupervised ML model using the unsupervised feature information, and deploys the local supervised and unsupervised ML models; determining when a detection performance difference between the local supervised and unsupervised ML models reaches a threshold; identifies a proposed change to the supervised or unsupervised feature information; deploys the proposed change on one client device; responsive to determining the proposed change improves the detection performance of that client device, communicates the proposed change to a sampled set of client devices; and responsive to determining the proposed change improves the detection performance of a majority of the sampled set, communicates the proposed change to the server.

MULTI-FACTOR AUTHENTICATION IN ENDPOINT DETECTION AND RESPONSE (17808188)

Main Inventor

Rosa M. Bolger

Brief explanation

The patent application describes techniques for mitigating cybersecurity incidents in a networked environment. Here are the key points:

• The techniques involve using an Endpoint Detection and Response (EDR) function to detect a security incident on a specific endpoint in a network.
• Once a security incident is detected, the techniques identify the administrator of the affected endpoint.
• A process requiring Multi-Factor Authentication (MFA) is initiated for the identified administrator.
• This is done by sending a push notification to a second device associated with the administrator.
• The administrator responds to the push notification, providing the required authentication.
• The EDR function then characterizes the maliciousness of the security incident based on the administrator's response.

Potential applications of this technology:

• Enhancing cybersecurity incident response in networked environments.
• Strengthening authentication processes for administrators.
• Improving the accuracy of characterizing the severity of security incidents.

Problems solved by this technology:

• Promptly detecting and mitigating cybersecurity incidents.
• Ensuring that only authorized administrators can respond to security incidents.
• Providing a more accurate assessment of the maliciousness of security incidents.

Benefits of this technology:

• Improved security by quickly detecting and responding to incidents.
• Enhanced protection against unauthorized access and malicious activities.
• More efficient and effective incident response by involving the appropriate administrator.

Abstract

Described are techniques for cybersecurity incident mitigation. The techniques include detecting, by an Endpoint Detection and Response (EDR) function in a networked environment comprising a plurality of endpoints, a security incident on a first endpoint of the plurality of endpoints. The techniques further include identifying an administrator of the first endpoint and initiating a process requiring Multi-Factor Authentication (MFA) associated with the administrator of the first endpoint by transmitting a push notification to a second device associated with the administrator and receiving a response to the push notification from the second device. The techniques further include characterizing, by the EDR function, a maliciousness of the security incident based on the response.

DISTRIBUTED MULTI-ACCESS EDGE SERVICE DELIVERY (18458410)

Main Inventor

Dharmendra MISRA

Brief explanation

The abstract of this patent application describes methods, computer program products, and systems for examining and joining capability data of a shared ledger data structure, specifically for service capability applications. Here is a simplified explanation of the abstract:

• The patent application presents methods, computer program products, and systems for managing service capability applications within a shared ledger data structure.
• The capability data of the shared ledger specifies the capabilities of various service capability applications.
• The method involves examining the capability data to determine the capabilities of each service capability application.
• Based on the examination, a service capability application is joined to a service delivery application.
• Once joined, the service delivery application can run with the added capabilities of the service capability application.

Potential Applications:

• This technology can be applied in various industries where service capability applications need to be integrated into service delivery applications.
• It can be used in sectors such as finance, healthcare, logistics, and more, where different applications with specific capabilities need to work together seamlessly.

Problems Solved:

• The technology solves the problem of efficiently managing and integrating service capability applications into service delivery applications.
• It provides a streamlined process for examining and joining capability data, reducing manual effort and potential errors.

Benefits:

• The technology improves the efficiency and effectiveness of service delivery applications by incorporating the capabilities of service capability applications.
• It allows for better collaboration and integration between different applications, enhancing overall system performance.
• The streamlined process saves time and resources by automating the examination and joining of capability data.

Abstract

Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: examining capability data of a shared ledger data structure, wherein the capability data specifies one or more capability for respective ones of a plurality of service capability applications; joining a service capability application of the plurality of service capability applications to a service delivery application in dependence on the examining; and running the service delivery application, with the service capability application joined to the service delivery application.

DECOUPLING CAPACITOR INSIDE GATE CUT TRENCH (17808178)

Main Inventor

REINALDO VEGA

Brief explanation

The abstract describes an approach to creating a semiconductor device that includes a decoupling capacitor to separate two gates. The decoupling capacitor consists of a dielectric liner and a ferroelectric material. The device also includes two power rails connected to the decoupling capacitor through the two gates.

• The semiconductor device includes a decoupling capacitor that separates two gates.
• The decoupling capacitor consists of a dielectric liner and a ferroelectric material.
• The device has two power rails connected to the decoupling capacitor through the two gates.

Potential Applications

• This technology can be used in various semiconductor devices such as integrated circuits and transistors.
• It can improve the performance and reliability of these devices by reducing noise and improving power distribution.

Problems Solved

• The decoupling capacitor helps to separate the two gates, reducing interference and improving the overall performance of the semiconductor device.
• By providing a separate power rail through the decoupling capacitor, power distribution can be more efficient and stable.

Benefits

• Improved performance and reliability of semiconductor devices.
• Reduced noise and interference.
• More efficient and stable power distribution.

References

[Link to the patent application]

Abstract

An approach to forming a semiconductor device where the semiconductor device includes a first power rail that is connected to a decoupling capacitor by way of a first gate. The decoupling capacitor is also connected to a second gate. As such, the decoupling capacitor separates the first gate from the second gate. The decoupling capacitor may include a dielectric liner within a gate cut trench and a ferroelectric material over the dielectric liner. A second power rail may be connected to the decoupling capacitor by way of the second gate. The first gate and the second gate may be inline with respect thereto.

MRAM DEVICE WITH SELF-ALIGNED BOTTOM ELECTRODES (17808561)

Main Inventor

Oscar van der Straten

Brief explanation

The patent application describes a magnetic tunnel junction (MTJ) stack that includes a bottom electrode surrounded by an oxide layer. The bottom electrode and oxide layer are horizontally aligned. Additionally, a metal spacer below the bottom electrode is also surrounded by the oxide layer, with its upper and lower surfaces horizontally aligned with the oxide layer.

• The MTJ stack includes a bottom electrode surrounded by an oxide layer.
• The oxide layer is horizontally aligned with the upper surface of the bottom electrode.
• A metal spacer below the bottom electrode is also surrounded by the oxide layer.
• The upper and lower surfaces of the metal spacer are horizontally aligned with the oxide layer.
• A metal layer is formed on the metal spacer and dielectric using a high-temperature deposition process, causing oxidation of the metal layer.

Potential Applications:

• Magnetic tunnel junction (MTJ) stacks can be used in various electronic devices, such as magnetic random-access memory (MRAM) and magnetic sensors.
• The described structure and process can enhance the performance and reliability of MTJ stacks in these devices.

Problems Solved:

• The oxide layer surrounding the bottom electrode and metal spacer helps to improve the stability and performance of the MTJ stack.
• The horizontal alignment of the surfaces ensures proper contact and reduces potential issues related to misalignment.

Benefits:

• The described structure and process can lead to improved performance, stability, and reliability of MTJ stacks.
• The use of the oxide layer and horizontal alignment helps to minimize potential issues and enhance the overall functionality of electronic devices utilizing MTJ stacks.

Abstract

A magnetic tunnel junction (MTJ) stack, where a vertical side surface of a bottom electrode of the MTJ stack is surrounded by an oxide, where the bottom electrode and the oxide are horizontally aligned. A magnetic tunnel junction (MTJ) stack, where a vertical side surface of a bottom electrode of the MTJ stack and a metal spacer below the bottom electrode is surrounded by an oxide, where an upper surface of the bottom electrode is horizontally aligned with a horizontal upper surface of the oxide, where a lower surface of the metal spacer is horizontally aligned with a horizontal lower surface of the oxide. Forming a metal spacer above and vertically aligned with a lower metal line surrounded by a dielectric, and forming a metal layer on the metal spacer and dielectric with a high temperature deposition of the metal layer, where the metal layer oxidizes.

BEVELED MAGNETO-RESISTIVE RANDOM ACCESS MEMORY PILLAR STRUCTURE (17808642)

Main Inventor

Oscar van der Straten

Brief explanation

The abstract describes a memory device that consists of a magnetic tunnel junction pillar connected to a bottom electrode and a top electrode. The pillar is made up of multiple layers stacked vertically above the bottom electrode. Each layer, as well as the top and bottom electrodes, is formed at a specific bevel angle. The width of the bottommost portion of each layer is larger than the width of the topmost portion of the preceding layer. An encapsulation layer is present along the sidewalls of the top and bottom electrodes, as well as the layers in the pillar.

• The memory device includes a magnetic tunnel junction pillar connected to electrodes.
• The pillar is composed of stacked layers with varying widths.
• Each layer, as well as the electrodes, is formed at a specific bevel angle.
• An encapsulation layer is present along the sidewalls of the electrodes and layers.

Potential Applications

This technology can have various applications in the field of memory devices, including:

• Non-volatile memory devices
• Magnetic random-access memory (MRAM)
• Data storage devices

Problems Solved

The innovation addresses the following problems:

• Improving the performance and efficiency of memory devices
• Enhancing the stability and reliability of memory storage
• Enabling higher data storage capacity

Benefits

The technology offers several benefits, such as:

• Increased memory device performance
• Enhanced stability and reliability of data storage
• Higher data storage capacity
• Improved efficiency and energy consumption

Abstract

A memory device includes a magnetic tunnel junction pillar located between, and electrically connected to, a bottom electrode and a top electrode. The magnetic tunnel junction pillar is composed of a plurality of device layers vertically stacked above the bottom electrode. Each of the plurality of device layers, the top electrode, and the bottom electrode is formed at a first bevel angle. A bottommost portion of each of the plurality of device layers in the magnetic tunnel junction pillar has a width that is greater than a width of a topmost portion of each preceding device layer. An encapsulation layer is disposed along opposite sidewalls of the top electrode, opposite sidewalls of the bottom electrode, and opposite sidewalls of each of the plurality of device layers.