Microsoft Technology Licensing, LLC patent applications published on October 12th, 2023

Summary of the patent applications from Microsoft Technology Licensing, LLC on October 12th, 2023

Microsoft Technology Licensing, LLC has filed several patents recently in various technological fields. These patents cover a range of innovative techniques and systems, including haptic feedback for thin keyboards, coordinating user interface formats for different client devices, viewing messages from multiple chat threads in a single user interface, starting instant messaging chat sessions from email threads, monitoring the quality of optical communications networks, uninterruptible power supply systems, laser systems for producing multiple laser beams, adjusting user-selected locations on visual map programs using non-visual feedback, and neural optimizers for fitting problems.

Notable applications include:

• Techniques for providing quick haptic feedback on individual keys of a thin keyboard or keypad, simulating the tactile sensation of pressing a mechanical key.
• Coordinating user interface formats for different client devices, allowing each device to display different user interface formats and separate messages based on priority topics.
• Viewing messages from multiple chat threads in a single user interface, identifying related threads based on user input or system queries.
• Starting instant messaging chat sessions from email threads, extracting user identifiers and displaying presence information for each user.
• Monitoring the quality of optical communications networks using Fourier transforms to analyze waveforms and calculate the optical signal to noise ratio (OSNR) penalty.
• Uninterruptible power supply systems that provide continuous power in case of power outages or anomalies, with separate AC and DC power outputs.
• Laser systems that efficiently produce multiple laser beams of higher frequencies from a lower frequency beam.
• Adjusting user-selected locations on visual map programs using non-visual feedback, such as audio or tactile triggers.
• Neural optimizers for fitting problems, running quickly and efficiently without manual input and incorporating information from previous iterations to improve accuracy.

These patents and applications demonstrate Microsoft Technology Licensing, LLC's commitment to innovation and advancing technology in various fields.

Patent applications for Microsoft Technology Licensing, LLC on October 12th, 2023

HOMOPOLYMER PRIMERS FOR AMPLIFICATION OF POLYNUCLEOTIDES CREATED BY ENZYMATIC SYNTHESIS (18204920)

Inventor Yuan-Jyue CHEN

Brief explanation

This disclosure presents a method for selectively amplifying certain polynucleotides from a pool using a unique primer and a homopolymer primer. The polynucleotides are synthesized using an enzyme called terminal deoxynucleotide transferase (TdT), which efficiently creates homopolymer sequences. By using only one homopolymer primer instead of two unique primers, the complexity, time, and cost of synthesizing the polynucleotides are reduced. The unique primer allows for the identification of different groups of polynucleotides, enabling random access through polymerase chain reaction (PCR) amplification. Additionally, the polynucleotides may contain regions that encode digital data using specific nucleotide sequences.

Abstract

This disclosure describes a technique for performing random access in a pool of polynucleotides by using one unique primer and one homopolymer primer to selectively amplify some but not all of the polynucleotides in the pool. The polynucleotides are synthesized by a template independent polymerase such as terminal deoxynucleotide transferase (TdT) rather than by phosphoramidite synthesis. Enzymatic synthesis efficiently creates homopolymer sequences through unregulated synthesis. Use of one homopolymer primer instead of two unique primers decreases the complexity, time, and cost of synthesizing the polynucleotides. Use of a unique primer provides a sequence that can be varied to uniquely identify multiple different groups of polynucleotides. This enables random access by polymerase chain reaction (PCR) amplification while still benefiting from the efficiency of homopolymer synthesis. The polynucleotides may include payload regions that use a sequence of nucleotides to encode digital data.

Device Cooling (17714425)

Inventor Chien Lung YANG

Brief explanation

The abstract describes a system for cooling devices using air. It explains that the system includes a component that generates heat, which is placed inside a housing. The system also includes an assembly that separates the ambient air into warmer air, which is directed outside the housing, and cooler air, which is directed inside the housing to cool the heat generating component.

Abstract

The description relates to devices and air cooling of devices. One example can include a heat generating component positioned in a housing and a temperature-based gas separation assembly configured to receive ambient air and to separate the ambient air into warmer air that is directed outside the housing and cooler air that is directed inside the housing to cool the heat generating component.

EXTENDING APPLICATION LIFECYCLE MANAGEMENT TO USER-CREATED APPLICATION PLATFORM COMPONENTS (18336013)

Inventor Marcio COSTA, JR.

Brief explanation

This abstract describes a method for extending application lifecycle management (ALM) processes to user-created application platform components. The method involves generating two components within an application platform. The first component can be customized by indicating whether it is subject to localization, defining its layering, and specifying if it is protected from modification. The second component is then customized based on the first component and by defining a dependency on it. These components are deployed in a target environment with metadata that represents the customizations and enables ALM processes.

Abstract

The examples described herein extend application lifecycle management (ALM) processes (e.g., create, update, delete, retrieve, import, export, uninstall, publish) to user-created application platform components. First and second components are generated within an application platform. The first component is customized at least by indicating whether the first component is subject to localization, defining a layering of the first component, and indicating whether the first component is protected from downstream modification. The second component is customized in accordance with customizing the first component, and is further customized by defining a dependency of the second component on the first component. The components are deployed in a target environment with metadata representing the customizations and enabling the ALM processes.

REGISTER ALLOCATION HEURISTICS ORDER (17717941)

Inventor Kunal PATHAK

Brief explanation

The abstract discusses how the process of compilation can be improved by changing the order in which certain steps are performed. This order can be determined based on various factors, such as the target processor architecture, operating system, type of source code, and optimization goals. A genetic algorithm is used to efficiently generate and evaluate different orders of steps until the best one is found. The evaluation is based on either real or simulated performance.

Abstract

Compilation is supported and improved by varying the order of invocation of register allocation heuristics during code generation. A particular invocation order may be chosen based on one or more compilation scenario properties, such as a target processor architecture, a target operating system, a kind of source code being compiled, or optimization targets for the compiler or the generated code, or a mix thereof. Suitable heuristics invocation orders may be produced efficiently and effectively using a genetic algorithm that is adapted to make a population of invocation orders, select parents, create offspring, and assess invocation order fitness, until the population converges on optimal orders. Invocation order fitness assessments may be based on actual performance or simulated performance.

DIRECT SWAP CACHING WITH ZERO LINE OPTIMIZATIONS (17718920)

Inventor Ishwar AGARWAL

Brief explanation

The abstract describes a method for managing a computer system with two types of memory: near memory and far memory. When a request is made to read a block of data, the method analyzes the metadata associated with the block. The metadata contains information about whether the block is stored in the near memory or far memory, as well as whether the data in the block is all zeros. Instead of retrieving the data from the far memory, the method synthesizes the data and sends it to the requester. This approach helps optimize the system's performance by avoiding unnecessary data retrieval.

Abstract

Systems and methods related to direct swap caching with zero line optimizations are described. A method for managing a system having a near memory and a far memory comprises receiving a request from a requestor to read a block of data that is either stored in the near memory or the far memory. The method includes analyzing a metadata portion associated with the block of data, the metadata portion comprising: both (1) information concerning whether the near memory contains the block of data or whether the far memory contains the block of data and (2) information concerning whether a data portion associated with the block of data is all zeros. The method further includes instead of retrieving the data portion from the far memory, synthesizing the data portion corresponding to the block of data to generate a synthesized data portion and transmitting the synthesized data portion to the requestor.

CONFIDENTIAL COMPUTE ARCHITECTURE INTEGRATED WITH DIRECT SWAP CACHING (17716823)

Inventor Ishwar AGARWAL

Brief explanation

This abstract describes a confidential compute architecture that incorporates direct swap caching. It explains a method for managing near and far memory in a system. When it is determined that the far memory contains an encrypted version of a specific block of data, the method retrieves the encrypted version from the far memory. It then decrypts the data using a specific key that is exclusively used by a virtual machine associated with the system. The decrypted version of the data is then provided to the requester. Additionally, the method involves swapping out another block of data from the near memory to the far memory if there is an address conflict with the first block of data. This second block of data is also encrypted using a specific key that is exclusively used by another virtual machine associated with the system.

Abstract

Systems and methods for a confidential compute architecture integrated with direct swap caching are described. An example method for managing a near memory and a far memory includes, in response to determining that the far memory contains an encrypted version of a first block of data, retrieving from the far memory the encrypted version of the first block of data, decrypting the first block of data using a first key for exclusive use by a first virtual machine associated with the system, and providing a decrypted version of the first block of data to the requestor. The method further includes swapping out a second block of data having an address conflict with the first block of data from the near memory to the far memory, where the second block of data is encrypted using a second key for exclusive use by a second virtual machine associated with the system.

DISTRIBUTED APPLICATION EXECUTION FOR CLOUD COMPUTING (18331913)

Inventor Souren AGHAJANYAN

Brief explanation

This abstract describes a technique for cloud computing that involves distributing the execution of applications. When a command is received to launch an application, the technique determines where each component of the application should be executed based on the type of data it consumes. If a component needs to be executed in a local computing facility instead of a public one, a request is sent from the public facility to the local facility to execute that component there. Once authorized, the data from the locally executed component is requested and received without the public facility having direct access to the data source at the local facility.

Abstract

Cloud computing techniques utilizing distributed application execution are disclosed herein. One example technique includes receiving a command to launch an application, and in response, determining an execution location corresponding to a type of data consumed by individual components of the application. Upon determining that one of the components is to be executed in a local computing facility, the example technique includes transmitting, from a public computing facility to the local computing facility, a request to execute the one of the components in the local computing facility instead of the public computing facility. Upon being authorized by the local computing facility, data is requested and received from the one of the components executed at the local computing facility without having direct access from the public computing facility to a data source at the local computing facility.

PROFILING A PROGRAM BASED ON A COMBINATION OF DYNAMIC INSTRUMENTATION AND SAMPLING (17718129)

Inventor Patrick Lothian NELSON

Brief explanation

This abstract describes a computer system that can analyze and gather information about a program while it is running. The system uses a profiler to collect data about the program's execution, including details about specific functions within the program. It then combines this data with additional information about how many times each function was called or how long it took to execute. The system then creates a summary of this combined information, which can be reported and used for further analysis.

Abstract

A computer system is configured to profile a program during an execution of the program. Profiling the program includes obtaining, by a first profiler, a first set of information associated with the execution of the program, and obtaining a second set of information associated with one or more executions of the function. The second set of information includes at least a call count or an execution time of each of the one or more executions of the function. The computer system then identifies a subset of the first set of information that is associated with the one or more executions of the function, and creates aggregate information based on the subset of the first set of information and the second set of information. The aggregated information is then reported.

MEASURING AND IMPROVING INDEX QUALITY IN A DISTRUBUTED DATA SYSTEM (18325765)

Inventor Babatunde Micheal OKUTUBO

Brief explanation

This abstract describes a method for measuring and improving the quality of an index in a distributed data system. The method involves determining various quality metrics for each partition of the data system, which indicate the quality of specific properties of each partition. These quality metrics are then combined to generate an overall index quality score, which measures the performance of the index. This score is used to automatically identify inefficiencies in the index and determine appropriate maintenance actions to improve its performance. Additionally, each quality metric can be individually analyzed to identify the specific database property that is most impacting the index's performance.

Abstract

Embodiments described herein are directed to measuring and improving an index quality of a distributed data system. For example, various quality metrics are determined on a per partition basis of the distributed data system. Each of the quality metrics are indicative of a quality of a particular property of a partition. The quality metrics are aggregated to generate an overall index quality score, which provides a measure of the performance of the index. The index quality score is utilized to automatically detect an inefficiency of the index and automatically determine that certain index maintenance actions should be automatically performed to improve the performance of the index. Each quality metric may also be individually analyzed to determine which database property is affecting the performance of the index the most.

METHODS AND SYSTEMS FOR PERFORMING A VECTORIZED DELETE IN A DISTRIBUTED DATABASE SYSTEM (18334913)

Inventor Wangchao LE

Brief explanation

This abstract describes techniques for clustering delete targets in a distributed database system. The goal is to efficiently delete files by grouping them into clusters based on certain constraints. The process involves retrieving a file to be deleted from a set of delete targets, scanning existing clusters of files marked for deletion to find clusters with similar constraints, and either adding the file to an existing cluster or creating a new cluster if no suitable cluster is found. Additionally, a delta array is generated for each file in the new cluster, which represents data items in the file and indicates which data items should be deleted based on a deletion signal.

Abstract

Example aspects include techniques for clustering delete targets for vectorized deletion including retrieving, from a set of delete targets in a distributed database system, a file to be deleted, scanning existing clusters of files marked for deletion to identify at least one existing cluster of files having constraints corresponding to the file, based on identifying the at least one existing cluster of files, adding the file to the at least one existing cluster of files to create a new cluster of files, based on failing to identify the at least one existing cluster of files, creating the new cluster of files including the file, and generating, for each file in the new cluster of files and based on a deletion signal, a delta array including multiple bits representing data items in each file and indicating, based on bit value, target data items to be deleted from each file.

METHOD AND SYSTEM OF RETRIEVING ASSETS FROM PERSONALIZED ASSET LIBRARIES (17716653)

Inventor Ji LI

Brief explanation

The abstract describes a system and method for retrieving assets from a personalized asset library. This system receives a search query and encodes it into embedding representations using a trained machine learning model. It then compares these representations to a collection of asset representations and identifies the most relevant assets as search results. These asset representations are automatically generated without human labeling.

Abstract

A system and method and for retrieving assets from a personalized asset library includes receiving a search query for searching for assets in one or more asset libraries, the one or more asset libraries including a personalized asset library; encoding the search query into embedding representations via a trained query representation machine-learning (ML) model; comparing, via a matching unit, the query embedding representations to a plurality of asset representations, each of the plurality of asset representations being a representation of one of the plurality of candidate assets; identifying, based on the comparison, at least one of the plurality of the candidate assets as a search result for the search query; and providing the identified plurality of candidate assets for display as the search result. The plurality of asset representations for the one or more assets in the personalized content library are generated automatically without human labeling.

Intelligently Identifying and Grouping Relevant Files and Providing an Event Representation for Files (18335221)

Inventor Madeline Schuster KLEINER

Brief explanation

The abstract describes a method and system for organizing files based on their relevance to a user. It involves retrieving user-specific data, analyzing it to determine a parameter related to file relevance, and then identifying and organizing relevant files into groups. The system also provides display data related to these groups of relevant files.

Abstract

A method and system for intelligently organizing one or more groups of relevant files may include retrieving a user data signal including user-specific data, analyzing the user data signal to identify a parameter relating to file relevance, identifying one or more relevant files in a storage medium based on the parameter at least one of a user category property, a lifecycle stage property, a relevant activity property, or an activity level property of one or more files in the storage medium, organizing the one or more relevant file into the one or more groups of relevant files, and providing for display data relating to the one or more groups of relevant files.

SYSTEMS AND METHODS FOR PROVIDING FEEDBACK TO WEBPAGE OWNERS (18031829)

Inventor Rahul Kumar GUPTA

Brief explanation

The abstract describes a client device with a browser application that allows users to access webpages over a network. The device has instructions stored on its storage device that, when executed by its processor, enable the browser application to be displayed on the device's screen. The browser application includes a feedback tool and can communicate with a server to access webpages. The webpages have metadata, including contact information for the webpage owner. When the feedback tool is selected, feedback information is transmitted to the webpage owner using the contact information from the webpage metadata.

Abstract

A client device has a browser application to present webpages, accessed via a network, to a user. The client device has a storage device with instructions stored thereon that, when executed by a processor of the client device, cause the client device to present a browser application on a client device display, wherein a graphical user interface (GUI) of the browser application includes a feedback tool and communicate with a server via the network to access a webpage having a page head including webpage metadata. The method further includes identifying a webpage owner contact information from the webpage metadata and, in response to selection of the feedback tool, transmitting feedback information to the webpage owner based on the webpage owner contact information.

AGENTLESS EXTRACTION OF CLOUD CREDENTIALS (17715749)

Inventor Josef WEIZMAN

Brief explanation

This abstract describes a computing system that can access a cloud storage and create a copy of some or all of the data stored there. The system then searches this copy for a specific data pattern associated with a credential. If the pattern is found, the system extracts an identifier and the level of access granted to that identifier. This access level is typically related to accessing a specific resource in the cloud. Finally, the system takes steps to reduce the risk of exposing the credential, such as notifying the owner of the cloud resource, deleting the credential, or modifying or revoking the access level associated with the credential.

Abstract

A computing system is configured to access a cloud storage and make a copy of at least a portion of the cloud storage. The copy of the at least portion of the cloud storage to search for a data pattern associated with a credential. In response to finding the data pattern associated with the credential, the computing system extracts an identifier associated with the credential and a scope of permission that the identifier is granted to. The scope of permission is associated with a permission to access a cloud resource. Finally, a risk of potential exposure of the credential is mitigated, such as (but not limited to) notifying an owner of the cloud resource, deleting the credential from the cloud, or modifying or revoking the scope of permission associated with the credential.

AGENT-BASED EXTRACTION OF CLOUD CREDENTIALS (17715755)

Inventor Josef WEIZMAN

Brief explanation

This abstract describes a computing system that installs an agent on a cloud consumer's computer. The agent scans the computer's storage for a specific data pattern associated with a credential. If the data pattern is found, the agent sends it to the computing system. The computing system then extracts an identifier from the data pattern, determines the level of permission associated with the identifier, and takes steps to reduce the risk of the credential being exposed.

Abstract

A computing system is configured to cause an agent to be installed at a cloud consumer computing system. The cloud consumer computing system is configured to access a cloud service. The agent is configured to scan at least a portion of storage of the cloud consumer computing system for a data pattern associated with a credential. In response to finding the data pattern associated with the credential, the agent sends the data pattern to the computing system. In response to receiving the data pattern, the computing system is configured to extract an identifier associated with the credential based on the data pattern, identify a scope of permission to which the identifier is granted, and mitigate a risk of potential exposure of the credential.

DETECTING AN ALGORITHMIC ATTACK AGAINST A HOSTED AI SYSTEM BASED ON INPUTS AND OUTPUTS OF THE HOSTED AI SYSTEM (17715014)

Inventor Hyrum Spencer ANDERSON

Brief explanation

The abstract describes techniques for detecting algorithmic attacks on a hosted artificial intelligence (AI) system. These attacks are identified by analyzing the inputs (queries) and outputs of the AI system. 

In the first example, a feature-based classifier model is used to generate a classification score. This score is based on numerical representations of the queries and outputs. If the classification score is equal to or greater than a certain threshold, an algorithmic attack is detected.

In the second example, a transformer-based model is used to generate a vector. This vector is created by providing a multivariate time series, which includes attributes of the inputs and outputs, as input to the transformer-based model. An algorithmic attack is detected if the distance between the generated vector and a reference vector is less than or equal to a certain threshold.

Abstract

Techniques are described herein that are capable of detecting an algorithmic attack against a hosted artificial intelligence (AI) system based on inputs (e.g., queries) and outputs of the hosted AI system. In a first example, a feature-based classifier model is used to generate a classification score based on features that are derived from numerical representations of the queries and the outputs, and an algorithmic attack is detected based on the classification score being greater than or equal to a score threshold. In a second example, a transformer-based model is used to generate a vector by providing a multivariate time series, which is based on attribute(s) of the inputs and attribute(s) of the outputs, as an input to the transformer-based model, and an algorithmic attack is detected based on a distance between the vector and a point corresponding to a reference vector being less than or equal to a distance threshold.

ANOMALOUS ACTIVITY DETECTION IN CONTAINER IMAGES (17716741)

Inventor Dotan Patrich

Brief explanation

The abstract describes an apparatus that uses a processor and memory to identify and analyze container images. Each container image is assigned a behavioral profile, and the processor checks if the activities in a specific container image comply with the corresponding activities in its assigned profile. If any activity in the container image deviates from the profile, the processor determines it as an anomalous activity and generates an alert.

Abstract

According to examples, an apparatus may include a processor and a memory on which is stored machine-readable instructions that may cause the processor to identify a container image group of a plurality of container image groups to which a subject container image corresponds, in which each of the plurality of container image groups is assigned a respective behavioral profile. The processor may also determine whether activities in the subject container image comply with corresponding activities identified in the behavioral profile of the identified container image group. Based on a determination that at least one activity in the subject container image fails to comply with a corresponding at least one activity identified in the behavioral profile of the identified container image group, the processor may determine that the subject container image includes an anomalous activity and output an alert indicating that the subject container image includes an anomalous activity.

TABLE COLUMN OPERATIONS FOR SPREADSHEETS (17833247)

Inventor Advait SARKAR

Brief explanation

The present technology is about improving spreadsheet tables. It introduces systems and methods to make it easier to apply certain operations to entire columns in a table. For example, when a cell event occurs in a spreadsheet, the system detects it and identifies the table column that the cell belongs to. It then applies specific instructions, called metadata, to the cell based on the configuration of that column. This simplifies the process of applying changes across a whole column in a table.

Abstract

Various embodiments of the present technology generally relate to enhancements to spreadsheet tables. More specifically, embodiments of the present technology include systems and methods for incorporating various table column operations into table columns headers, thereby simplifying the act of propagating a desired effect across a range that exactly matches the table column. In an embodiment, a spreadsheet application detects a cell event in a spreadsheet, wherein the cell event affects one or more cells of the spreadsheet, determines that a cell of the one or more cells is inside a range that makes up a table in the spreadsheet, identifies a table column of the table to which the cell belongs, identifies metadata for the table column, wherein the metadata for the table column comprises column configuration instructions, and applies the column configuration instructions to the cell.

GROUNDED TEXT GENERATION (18334065)

Inventor Michel GALLEY

Brief explanation

The abstract describes a framework that can generate text based on input text. It consists of a machine learning model, a grounding interface, and a control interface. The machine learning model is trained to generate text and can access additional information related to the input text through the grounding interface. It can also recognize a control signal through the control interface. The model incorporates the information from the grounding source into the generated text and adjusts the focus of the generated text based on the control signal.

Abstract

A controllable grounded response generation framework includes a machine learning model, a grounding interface, and a control interface. The machine learning model is trained to output computer-generated text based on input text. The grounding interface is useable by the machine learning model to access a grounding source including information related to the input text. The control interface is useable by the machine learning model to recognize a control signal. The machine learning model is configured to include information from the grounding source in the computer-generated text and focus the computer-generated text based on the control signal.

ROBUSTNESS AGAINST MANIPULATIONS IN MACHINE LEARNING (18333487)

Inventor Cheng ZHANG

Brief explanation

The abstract describes a method that involves receiving data points with different feature values. These data points represent observations of a ground truth, which is represented by the feature vector. The method also involves learning parameters of a machine-learning model based on the observed data points. This model includes statistical models that capture the causal relationship between the feature vector, a latent vector, a classification, and a manipulation vector. The manipulation vector represents the effect of potential manipulations between the ground truth and its observation through the feature vector. The learning process involves mapping between the feature vector, latent vector, classification, and manipulation vector.

Abstract

A method comprising: receiving observed data points each comprising a vector of feature values, wherein for each data point, the respective feature values are values of different features of a feature vector. Each observed data point represents a respective observation of a ground truth as observed in the form of the respective values of the feature vector. The method further comprises learning parameters of a machine-learning model based on the observed data points. The machine-learning model comprises one or more statistical models arranged to model a causal relationship between the feature vector and a latent vector, a classification, and a manipulation vector. The manipulation vector represents an effect of potential manipulations occurring between the ground truth and the observation thereof as observed via the feature vector. The learning comprises learning parameters of the one or more statistical models to map between the feature vector, latent vector, classification and manipulation vector.

DIFFERENTIAL RECURRENT NEURAL NETWORK (18209989)

Inventor Patrice SIMARD

Brief explanation

The abstract describes a type of neural network called a differential recurrent neural network (RNN). This network is designed to handle dependencies between inputs that occur at different points in time. It does this by using recurrent loops to store and update states without affecting the training process. 

The differential RNN consists of three main components: a state component for storing states, a trainable transition and differential non-linearity component that includes a neural network, and a trainable OUT component that also includes a neural network for post-processing the current states.

The transition and differential non-linearity component takes as input the previous stored states from the state component and an input vector. It then produces positive and negative contribution vectors, which are used to create a state contribution vector. This state contribution vector is then input into the state component to generate a set of current states.

In one implementation, the current states are directly outputted. In another implementation, the current states are further processed by the trainable OUT component before being outputted.

Overall, this differential RNN allows for the handling of long-term dependencies in a neural network by effectively storing and updating states without interfering with the training process.

Abstract

A differential recurrent neural network (RNN) is described that handles dependencies that go arbitrarily far in time by allowing the network system to store states using recurrent loops without adversely affecting training. The differential RNN includes a state component for storing states, and a trainable transition and differential non-linearity component which includes a neural network. The trainable transition and differential non-linearity component takes as input, an output of the previous stored states from the state component along with an input vector, and produces positive and negative contribution vectors which are employed to produce a state contribution vector. The state contribution vector is input into the state component to create a set of current states. In one implementation, the current states are simply output. In another implementation, the differential RNN includes a trainable OUT component which includes a neural network that performs post-processing on the current states before outputting them.

GENERATING AND PROCESSING CONTEXTUALIZED GROUP DATA (17704801)

Inventor Aleksander TORSTENSEN

Brief explanation

This technology allows for better control and presentation of group data on computing devices, leading to improved computing applications, electronic communications, and user experiences. The system can determine or infer a group of people based on user activity. It generates and personalizes group data for the user based on their context or intent. When the user interacts with a group or group member, relevant aspects of the contextualized group data are provided. The group data can be presented in a formatted way based on the user's context and can be used by computing applications or services to enhance the user's computing experience.

Abstract

Technology is disclosed for controlling the processing and presentation of group data on computing devices to provide improved computing applications, electronic communications, and user computing experiences. Data indicating a group of people may be programmatically determined or inferred, such as from the user activity of a particular user. Group data characterizing the group of people may be generated and contextualized or personalized for the user based on a user context or intent. Aspects of the contextualized group data may be provided to the user in response to an interaction with an indication of the group or group member. The group data may be provided as a set of group information items, and further may be formatted or assembled for presentation based on the user context, and/or may be consumed by a computing application or service to provide an improved user computing experience.

MEETING RECURRENCE PLACEMENT OPERATOR (18026808)

Inventor Warren David Johnson III

Brief explanation

The abstract describes a system and method for allocating a recurring resource in an application to different users' computing devices. The system determines the frequency and range of the recurring resource, and then calculates a period based on this information. It accesses user resource data for each period and calculates a resource availability score. If the score for a particular period is below a certain threshold, the system allocates the recurring resource to instances of the period where the score exceeds the threshold.

Abstract

A system and method for allocating a recurring resource is described. The system receives a request to allocate the recurring resource of an application to computing devices associated with one or more users. The system identifies a cadence of the recurring resource and a range based on the cadence. A period is determines based on the cadence and the range. The system accesses user resource data of the application for each period and iteratively computes a resource availability score for the period based on the corresponding user resource data. The system determines that the resource availability score of an instance of the period is less than a threshold value and allocates the recurring resource to instances of the period where the corresponding resource availability score exceeds the threshold value.

Camera Calibration (17717354)

Inventor Hongli DENG

Brief explanation

The abstract describes a method for enhancing user experiences with cameras by calibrating their height and tilt angle. The method uses two images of a user taken at different locations relative to the camera, both showing only the user's upper body. By estimating the distance between the two locations and analyzing the images, the camera's position can be adjusted without needing a full body image of the user.

Abstract

The description relates to cameras, and camera calibration for enhancing user experiences. One example can receive a first image of a user at a first location relative to a camera. The first image can include the user's upper body but does not include the user from head to toe. The example can receive a second image of the user at a second location relative to a camera. The second image can include the user's upper body but does not include the user from head to toe. The example can estimate a distance of the second location from the first location relative to the camera and calibrate a height and tilt angle of the camera from the first image, the second image, and the estimated distance and without a full body image of the user.

CONCURRENT HUMAN POSE ESTIMATES FOR VIRTUAL REPRESENTATION (17658809)

Inventor Thomas Joseph CASHMAN

Brief explanation

This abstract describes a method for creating virtual representations of human body poses. It involves using sensors to gather data about the positioning of different body parts. This data is then used to estimate the poses of a model representation and a target representation. The method also includes maintaining mapping constraints that relate the model representation to the target representation. A pose optimization machine is trained using training data that has accurate labels for the model representation. The final output is a virtual representation of the human user displayed with the target pose.

Abstract

A method for virtually representing human body poses includes receiving positioning data detailing parameters of one or more body parts of a human user based at least in part on input from one or more sensors. One or more mapping constraints are maintained that relate a model articulated representation to a target articulated representation. A model pose of the model articulated representation and a target pose of the target articulated representation are concurrently estimated based at least in part on the positioning data and the one or more mapping constraints. The previously-trained pose optimization machine is trained with training positioning data having ground truth labels for the model articulated representation. The target articulated representation is output for display with the target pose as a virtual representation of the human user.

SYSTEMS AND METHODS FOR USER-GENERATED CONTENT WITH MACHINE-GENERATED METADATA (17716879)

Inventor Donyea Durece COOLEY-WHITE

Brief explanation

This abstract describes a method for helping with the sharing of video information. It involves a server computer obtaining a comparison video from a social media platform and determining if there is an event in the video. The method also includes gathering social media metrics for the video and evaluating its description. If there are descriptors in the description that match the event in the video, they are recorded in an application module.

Abstract

A method of assisting video information sharing includes, at a server computer, obtaining a comparison video including a plurality of frames from a social media platform and determining a presence of at least one event in the comparison video. The method further includes obtaining social media metrics for the comparison video from the social media platform and evaluating a description of the comparison video. The method further includes identifying at least one descriptor in the description correlated to the at least event in the comparison video and recording the descriptor in an application module.

OPTIMIZATION-BASED PARAMETRIC MODEL FITTING VIA DEEP LEARNING (17719335)

Inventor Julien Pascal Christophe VALENTIN

Brief explanation

The abstract describes a neural optimizer that can be used for various fitting problems. It can run quickly and efficiently without the need for manual input, and it incorporates information from previous iterations to improve accuracy. The optimizer controls the learning rate of each parameter individually, ensuring robustness and fast convergence. It combines gradient descent with a method that rapidly reduces the fitting energy. Overall, the abstract highlights the capabilities and advantages of this neural optimizer.

Abstract

A neural optimizer is disclosed that is easily applicable to different fitting problems, can run at interactive rates without requiring significant efforts, does not require hand crafted priors, carries over information about previous iterations of the solve, controls the learning rate of each parameter independently for robustness and convergence speed, and combines updates from gradient descent and from a method capable of very quickly reducing the fitting energy. A neural fitter estimates the values of the parameters Θ by iteratively updating an initial estimate Θ.

LOCATION ADJUSTMENT WITH NON-VISUAL LOCATION FEEDBACK (17840156)

Inventor Melanie Jo KNEISEL

Brief explanation

The abstract describes devices, systems, and methods for adjusting a user-selected location on a visual map program using non-visual feedback. The method involves initially indicating the selected location, detecting an audio or tactile trigger from the user, and then adjusting the selected location on the map based on the trigger. This adjustment is accompanied by providing audio or tactile feedback to indicate the new location on the map.

Abstract

Generally discussed herein are devices, systems, and methods for location adjustment with non-visual location feedback. A method can include providing, by a visual map program, an indication of an initial selected location of a user-selected location, detecting, by data provided by an input component, an audio or tactile trigger by a user, responsive to the detecting, altering, by a first distance and a first direction, the initial selected location to a new location on the visual map program, and responsive to the altering, causing an output component to provide an audio or tactile indication of the new location.

GENERATING MULTIPLE BEAMS OF A HARMONIC FREQUENCY (17804368)

Inventor Richard John BLACK

Brief explanation

The abstract describes a laser system that efficiently produces multiple laser beams of a higher frequency (harmonic frequency) from a lower frequency (fundamental frequency) beam. The system includes a laser that emits the fundamental frequency beam, a first stage that converts the fundamental frequency beam into a first-stage harmonic frequency beam and a first-stage residual fundamental frequency beam, and a second stage that further converts the residual fundamental frequency beam into a second-stage harmonic frequency beam.

Abstract

Examples are disclosed that relate to efficiently producing multiple laser beams of a harmonic frequency from a fundamental frequency beam. One example provides a laser system comprising a laser configured to output a fundamental frequency beam, a first harmonic-generation stage, and a second harmonic-generation stage. The first harmonic-generation stage is configured to receive an input of the fundamental frequency beam from the laser, and output from the laser system a first-stage harmonic frequency beam and a first-stage residual fundamental frequency beam. The second harmonic-generation stage is configured to receive an input of the first-stage residual fundamental frequency beam, and to output from the laser system a second-stage harmonic frequency beam.

DUAL OUTPUT UNINTERRUPTIBLE POWER SUPPLY (18335095)

Inventor Banha SOK

Brief explanation

The abstract describes an uninterruptible power supply (UPS) that is designed to provide continuous power in case of a power outage or other power anomalies. The UPS has a rectifier that converts alternating current (AC) power into a usable form. It has two outputs - one that provides AC power and another that provides DC power. The AC power output is created by an inverter connected to the rectifier, while the DC power output is created by a battery backup and a stepdown converter connected to the rectifier. The DC power output is activated when a power anomaly is detected, ensuring that power is still available even during disruptions. This design increases power availability and uptime, making it a reliable solution for uninterrupted power supply.

Abstract

An uninterruptible power supply (UPS) includes a rectifier configured to receive alternating current (AC) power. The UPS further includes a first output connected to the rectifier through an inverter. The first output is configured to output an AC power supply. The UPS also includes a second output connected to the rectifier through a battery backup and a stepdown converter. The second output is configured to output a direct current (DC) power supply in response to a detected power anomaly condition, thereby providing extra redundancy that allows for increased power availability and uptime.

CHARACTERIZATION OF INTER-CHANNEL CROSSTALK IN AN OPTICAL NETWORK (18024599)

Inventor Yawei YIN

Brief explanation

This abstract describes a system for monitoring the quality of an optical communications network. The system uses Fourier transforms to analyze the waveforms of data channels and calculate the noise-to-signal ratio. Based on this ratio, an estimate of the optical signal to noise ratio (OSNR) penalty is made. If the OSNR penalty exceeds a certain threshold, a notification is generated.

Abstract

An optical communications network comprises optical data links comprising data channels. A time-domain sampled waveform of a selected data channel is obtained. The Fourier transform is applied to the time-domain sampled waveform of the selected data channel to generate a frequency-domain waveform of the selected data channel. Time-domain sampled waveforms of the selected data channels neighboring data channels are obtained. The Fourier transform is applied to the time-domain sampled waveforms of the neighboring data channels to generate frequency-domain waveforms of the neighboring data channels. The noise-to-signal ratio is calculated based on the frequency-domain waveforms. Based on the calculated noise-to-signal ratio, an optical signal to noise ratio (OSNR) penalty is estimated. A notification is generated when the OSNR penalty exceeds a predetermined threshold.

DATA FLOW CONTROL IN DISTRIBUTED COMPUTING SYSTEMS (17714514)

Inventor Ariane Belle Tsai

Brief explanation

Datacenters and distributed computing systems offer computing resources like processing power and data storage as services accessible through a computer network. These services are used by tenants, such as corporations or schools, who have multiple users or groups with their own websites. To make it easier for users to access their data, relevant information is stored in a dedicated network location called a shard. A shard can be a physical or logical storage location that contains emails, chats, documents, photos, videos, and other types of content that users can interact with.

Abstract

Datacenters or other large-scale distributed computing systems can provide computing resources such as processing power and data storage as computing services accessible to tenants via a computer network. A tenant, such as a corporation, school, or organization, can have multiple users or groups of users with corresponding websites. To facilitate ready access, data relevant to a user, group, or website of a tenant can be stored in a dedicated network location sometimes referred to as a shard. A shard can be a physical and/or logical storage location that contains emails, chats, instant messages, documents, photos, videos, or other types of content items with which the user, group, or website can interact.

SYSTEM AND METHOD FOR CONVERTING AN EMAIL THREAD TO AN INSTANT MESSAGE GROUP (18207509)

Inventor Afraz A. SIDDIQUI

Brief explanation

This abstract describes a system and method for starting an instant messaging chat session from an email thread. It explains that the system receives an email thread and extracts user identifiers from the emails. It then obtains presence information for each user identifier and displays them in a user interface window. Based on the presence information, a user control for instant messaging is displayed. If the user selects this control, an instant messaging chat session is initiated.

Abstract

Systems and methods for initiating an instant messaging chat session from an email thread are described. In examples, an email thread including at least one email is received and user identifiers form the at least one email are extracted. Presence information associated with each user identifier of the plurality of user identifiers is obtained and then ordered. The ordered plurality of user identifiers together with their presence information are displayed in a user interface window. A determination to display a user control associated with an instant messaging capability is based on the presence information associated with at least one user identifier of the plurality of user identifiers. Upon receiving an indication that the user control associated with the instant messaging chat capability is selected, an instant messaging chat session is initiated.

CONTROLLED DISPLAY OF RELATED MESSAGE THREADS (18191739)

Inventor Amer Aref HASSAN

Brief explanation

The system described in this abstract allows users to view messages from multiple chat threads in a single user interface. It can identify related threads based on user input or system queries, and only searches for threads that have at least one user in common with the initial thread. The system displays the related threads without merging their messages, and uses a data structure to control the display of each thread.

Abstract

The techniques disclosed herein provide a system that can identify two or more independently managed message threads and modify a user interface to enable a common user of the threads to view messages of both threads. For instance, if a user is participating in a first chat session that is displayed within a first message thread, the system can identify other threads that are related to the first thread. The related threads can be identified by a user input or by a query performed by a system. In some embodiments, the system can only search for threads that have at least one common user to the first thread. The system can provide this display of two related threads while linking the threads but not merging the messages of the threads. A data structure can be maintained to control the display of each thread.

COORDINATION OF MESSAGE THREAD GROUPINGS ACROSS DEVICES OF A COMMUNICATION SYSTEM (18199327)

Inventor Amer Aref HASSAN

Brief explanation

The abstract describes a system that can coordinate user interface formats for different client devices. Each device is managed by a data structure that defines thread containers, which contain threads and topics. Each thread can have unique groups of messages. This allows each client device to display different user interface formats, showing different threads for each user. Users can split a parent thread into multiple child threads, each associated with unique topics. Messages from the parent thread can be associated with each child thread. This system provides more control over user interface formats, allowing users to separate messages based on priority topics.

Abstract

A system that can coordinate user interface formats having customized threads comprising messages from a common thread. Each client device can be managed by a data structure that defines thread containers, where individual thread containers define threads and topics for each thread. Each thread can also be associated with unique groups of messages. The data structure can cause each client computer to display different user interface formats, with each format showing different threads for each user. Users engaged in a chat on a parent thread can split the parent thread into multiple child threads, that can each be associated with unique topics. The system can then associate individual messages of the parent thread with each child thread. The system provides more granular control of user interface formats to allow users to separate messages of high-priority topics.

CONTROLLER-LESS QUICK TACTILE FEEDBACK KEYBOARD (18209906)

Inventor Masaaki FUKUMOTO

Brief explanation

This abstract describes techniques for providing quick haptic feedback on individual keys of a thin keyboard or keypad, without the need for a controller. The haptic feedback is designed to simulate the tactile sensation of pressing a mechanical key, giving the user the perception that they have actuated a physically movable key.

Abstract

In some examples, techniques are provided for quick haptic feedback, without the use of a controller, which is local to individual, non-actuating keys, such as keys of a thin keyboard or keypad. The haptic feedback may be in the form of a simulated “key-click” feedback for an individual key that is pressed by a user such that the finger used to press the key feels the tactile sensation. The haptic feedback mimics the tactile sensation of a mechanical key (e.g., buckling spring, pop-dome key switch) to give a user the perception that they have actuated a mechanically movable key.