GOOGLE LLC patent applications on April 25th, 2024
Patent Applications by GOOGLE LLC on April 25th, 2024
GOOGLE LLC: 44 patent applications
GOOGLE LLC has applied for patents in the areas of G06N20/00 (8), G06N3/08 (8), G10L15/26 (8), G10L15/22 (7), G10L15/30 (6)
With keywords such as: data, based, user, device, training, network, video, item, include, and input in patent application abstracts.
Patent Applications by GOOGLE LLC
20240131695.DEEP REINFORCEMENT LEARNING FOR ROBOTIC MANIPULATION_simplified_abstract_(google llc)
Inventor(s): Sergey Levine of Berkeley CA (US) for google llc, Ethan Holly of San Francisco CA (US) for google llc, Shixiang Gu of Mountain View CA (US) for google llc, Timothy Lillicrap of London (GB) for google llc
IPC Code(s): B25J9/16, G05B13/02, G05B19/042, G06N3/008, G06N3/045, G06N3/08
Abstract: implementations utilize deep reinforcement learning to train a policy neural network that parameterizes a policy for determining a robotic action based on a current state. some of those implementations collect experience data from multiple robots that operate simultaneously. each robot generates instances of experience data during iterative performance of episodes that are each explorations of performing a task, and that are each guided based on the policy network and the current policy parameters for the policy network during the episode. the collected experience data is generated during the episodes and is used to train the policy network by iteratively updating policy parameters of the policy network based on a batch of collected experience data. further, prior to performance of each of a plurality of episodes performed by the robots, the current updated policy parameters can be provided (or retrieved) for utilization in performance of the episode.
Inventor(s): Daniel Joseph Filip of San Jose CA (US) for google llc, Zhen Yang of San Jose CA (US) for google llc
IPC Code(s): G01C21/36, G01C21/28, G06F18/214, G06T7/80, G06V20/10, H04N23/698
Abstract: the present disclosure is directed to determining geographic orientation based at least in part on imagery. in particular, the methods and systems of the present disclosure can: receive data generated by a camera and representing imagery that includes at least a portion of a physical real-world environment comprising the camera and a travelway; and determine, based at least in part on the data and a machine-learning model, a geographic orientation of the camera with respect to the travelway.
Inventor(s): Wei Jin of Saratoga CA (US) for google llc, Joseph Daniel Lowney of Tucson AZ (US) for google llc, Lu Tian of Palo Alto CA (US) for google llc, Qinglan Huang of Mountain View CA (US) for google llc, Thomas Mercier of Weston FL (US) for google llc
IPC Code(s): G02B27/01, G02B6/34
Abstract: an augmented-reality (ar) eyewear display utilizes an optical waveguide having multi-layered optical gratings in a repeating arrangement. the optical gratings include varying depths, slope angles, lengths, and/or widths in order to tune the gratings to provide an improved ar eyewear display. by using the different configurations of two-dimensional or three-dimensional gratings disclosed herein in a waveguide of an ar eyewear display, optical characteristics of the waveguide are optimized to provide, e.g., high resolution and/or contrast, high display uniformity, high input coupling efficiency, and/or high output coupling efficiency. accordingly, in some embodiments, aspects of the present disclosure enable lower-power ar eyewear displays to produce the same quality of display of a higher-power conventional ar eyewear display waveguide.
Inventor(s): Lu Tian of Palo Alto CA (US) for google llc, Wei Jin of Saratoga CA (US) for google llc, Thomas Mercier of Weston FL (US) for google llc
IPC Code(s): G03F7/00, G02B6/34, G02B27/01, G03F7/40
Abstract: a method for fabricating a working stamp for forming surface gratings in a waveguide workpiece includes performing a series of step lithography processes using a series of master stamps so as to form a working stamp having a first surface having a plurality of surface gratings formed therein. each step lithography process includes pressing a master stamp of the series of master stamps into a material layer of a working stamp workpiece at a corresponding region of the material layer, the master stamp having a plurality of surface grating patterns formed thereon. each step lithography process further includes applying ultraviolet light to the corresponding region to locally cure the material layer at the corresponding region and detaching the master stamp from the material of the working stamp after applying the ultraviolet light.
20240134462.CONFIDENCE-BASED APPLICATION-SPECIFIC USER INTERACTIONS_simplified_abstract_(google llc)
Inventor(s): Ashton Udall of San Bruno CA (US) for google llc, Andrew Christopher Felch of Palo Alto CA (US) for google llc, James Paul Tobin of San Francisco CA (US) for google llc
IPC Code(s): G06F3/01, G01V1/00, G06V40/20
Abstract: this application is directed to a method for controlling user experience (ux) operations on an electronic device that executes an application. a touchless ux operation associated with the application has an initiation condition including at least detection of a presence and a gesture in a required proximity range with a required confidence level. the electronic device then determines from a first sensor signal the proximity of the presence with respect to the electronic device. in accordance with a determination that the determined proximity is in the required proximity range, the electronic device determines from a second sensor signal a gesture associated with the proximity of the presence and an associated confidence level of the determination of the gesture. in accordance with a determination that the determined gesture and associated confidence level satisfy the initiation condition, the electronic device initializes the touchless ux operation associated with the application.
20240134708.Bin Packing_simplified_abstract_(google llc)
Inventor(s): Md Ehtesamul Haque of Santa Clara CA (US) for google llc, Thomas John Chestna of Middleborough MA (US) for google llc, Samuel Justin Smith of Mountain View CA (US) for google llc, Pedro Daniel Valenzuela Salvatierra of Santa Clara CA (US) for google llc, Olivier Robert Sevin of Winston-Salem NC (US) for google llc
IPC Code(s): G06F9/50, G06F9/455
Abstract: a system and method for assigning a workload to one of a plurality of candidate host machines of a computing environment. the method may include receiving a request to schedule a workload, selecting a virtual machine type for executing the workload, for each candidate host machine of the plurality of candidate host machines, determining an expected waste score indicating a likelihood of resources at the candidate host machine remaining unused if the virtual machine type is assigned to the candidate host machine, selecting the candidate host machine for which the expected waste score is the lowest, and assigning the workload to the selected candidate host machine.
Inventor(s): Madhav Datt of Mountain View CA (US) for google llc, Sukriti Ramesh of Bengaluru (IN) for google llc
IPC Code(s): G06F16/242, G06N20/00
Abstract: provided are computing systems, methods, and platforms for generating training and testing data for machine-learning models. the operations can include receiving signal extraction information that has instructions to query a data store. additionally, the operations can include accessing, using structured query language (sql) code generated based on the signal extraction information, raw data from the data store. moreover, the operations can include processing the raw data using signal configuration information to generate a plurality of signals. the signal configuration information can have instructions on how to generate the plurality of signals from the raw data. furthermore, the operations can include joining, using sql code, the plurality of signals with a first label source to generate training data and testing data. subsequently, the operations can include processing the training data and the testing data to generate the input data. the input data being an ingestible for a machine-learning pipeline.
20240134893.STREAMING REAL-TIME DIALOG MANAGEMENT_simplified_abstract_(google llc)
Inventor(s): David Elson of Port Washington NY (US) for google llc, Christa Wimberley of Mountain View CA (US) for google llc, Benjamin Ross of Mountain View CA (US) for google llc, David Eisenberg of New York NY (US) for google llc, Sudeep Gandhe of Mountain View CA (US) for google llc, Kevin Chavez of Mountain View CA (US) for google llc, Raj Agarwal of Mountain View CA (US) for google llc
IPC Code(s): G06F16/332, G06F16/00, G06F16/33, G06F40/30, G06N20/00, G06Q10/10, G10L15/22
Abstract: systems and methods provides for dialog management in real-time rather than turn taking. an example method included generating first candidate responses to triggering event. the triggering event may be receipt of a live stream chunk for the dialog or receipt of a backend response to a previous backend request for a dialog shema. the method also includes updating a list of candidate responses that are accepted or pending with at least on of the first candidate responses, and determining, for the triggering event, whether the list of candidate responses includes a candidate response that has a confidence score that meets a triggering threshold. the method also includes waiting for a next triggering event without providing a candidate response when the list does not include a candidate response that has a confidence score that meets the triggering threshold.
Inventor(s): Guannan Zhang of Shanghai (CN) for google llc, Yiling Zhang of Shanghai (CN) for google llc
IPC Code(s): G06F16/955, G06F16/23
Abstract: systems and methods for automatically associating content characteristics to a third-party content are provided. a uniform resource locator identifying a resource is received from a content provider. the content is rendered to produce an object tree. a first node of the object tree is determined, where the first node represents a content slot. a second node of the object tree proximate to the first node is identified. the second node has a content characteristic, which is extracted. the extracted content characteristic is associated with the first node and stored.
20240134980.IDENTIFY MALICIOUS SOFTWARE_simplified_abstract_(google llc)
Inventor(s): Richard Cannings of Santa Cruz CA (US) for google llc, Sai Deep Tetali of Mountain View CA (US) for google llc, Mo Yu of Mountain View CA (US) for google llc, Salvador Mandujano of San Jose CA (US) for google llc
IPC Code(s): G06F21/56, G06F21/52, G06N3/04, G06N3/08
Abstract: a method for identifying malicious software includes receiving and executing a software application, identifying a plurality of uniform resource identifiers the software application interacts with during execution of the software application, and generating a vector representation for the software application using a feed-forward neural network configured to receive the plurality of uniform resource identifiers as feature inputs. the method also includes determining similarity scores for a pool of training applications, each similarity score associated with a corresponding training application and indicating a level of similarity between the vector representation for the software application and a respective vector representation for the corresponding training application. the method also includes flagging the software application as belonging to a potentially harmful application category when one or more of the training applications have similarity scores that satisfy a similarity threshold and include a potentially harmful application label.
20240134986.Securely Provisioning Secrets in Authentication Devices_simplified_abstract_(google llc)
Inventor(s): Vidya Satyamsetti of Bothell WA (US) for google llc
IPC Code(s): G06F21/57, G06F21/44
Abstract: the present disclosure provides for increased security of root of trust (rot) chips by preventing malicious provisioning. unique device secrets (uds) can only be provisioned securely by trustworthy hardware or trustworthy firmware. entities other than the trustworthy hardware and trustworthy firmware do not have access to a composite device identifier (cdi) generated using the uds and firmware measurements.
20240135042.Using Memory Protection Data_simplified_abstract_(google llc)
Inventor(s): Yanru Li of San Diego CA (US) for google llc, Deepti Vijayalakshmi Sriramagiri of San Jose CA (US) for google llc
IPC Code(s): G06F21/64
Abstract: the present disclosure describes techniques and apparatuses that are directed to using memory protection data within a computing device. techniques include allocating regions of a memory for storing application data and protection data. techniques also include creating a bitmap having bit values corresponding to memory blocks within the allocated regions. the one or more bit values can be indicative of whether application data and/or protection data are present in a memory block. the techniques and apparatuses can enable memory protection, such as memory security (e.g., encryption) and memory safety (e.g., error correction code (ecc) usage), to be efficiently used while permitting discontiguous memory allocations and without substantial operating system modification.
20240135117.METHOD FOR SPEECH-TO-SPEECH CONVERSION_simplified_abstract_(google llc)
Inventor(s): Oleg RYBAKOV of Redmond WA (US) for google llc, Fadi BIADSY of Sandyston NJ (US) for google llc
IPC Code(s): G06F40/58, G10L15/02, G10L15/30, G10L19/032
Abstract: the present disclosure relates to a streaming speech-to-speech conversion model, where an encoder runs in real time while a user is speaking, then after the speaking stops, a decoder generates output audio in real time. a streaming-based approach produces an acceptable delay with minimal loss in conversion quality when compared to other non-streaming server-based models. a hybrid model approach for combines look-ahead in the encoder and a non-causal stacker with non-causal self-attention.
Inventor(s): Alexander James Faaborg of Mountain View CA (US) for google llc, Brett Aladdin Barros of San Mateo CA (US) for google llc
IPC Code(s): G06K7/14, G06K19/06, G06V10/20
Abstract: the present disclosure relates generally to the processing of machine-readable visual encodings in view of contextual information. one embodiment of aspects of the present disclosure comprises obtaining image data descriptive of a scene that includes a machine-readable visual encoding; processing the image data with a first recognition system configured to recognize the machine-readable visual encoding; processing the image data with a second, different recognition system configured to recognize a surrounding portion of the scene that surrounds the machine-readable visual encoding; identifying a stored reference associated with the machine-readable visual encoding based at least in part on one or more first outputs generated by the first recognition system based on the image data and based at least in part on one or more second outputs generated by the second recognition system based on the image data; and performing one or more actions responsive to identification of the stored reference.
Inventor(s): Madhav Datt of Mountain View CA (US) for google llc, Surabhi Choudhary of Chennai (IN) for google llc, Nikhil Shirish Ketkar of Bengaluru (IN) for google llc
IPC Code(s): G06N3/048, G06N3/08
Abstract: provided are computing systems, methods, and platforms for a discrete-valued output classification. the operations can include obtaining a candidate threshold value for a first slice in a plurality of data slices. additionally, the operations can include calculating, using a candidate machine-learned model and the candidate threshold value, a first performance value associated with a first risk tolerance value. moreover, the operations can include determining, based on the first performance value, that a safeguard criterion for the first slice has not been satisfied. in response to the determination that the safeguard criterion for the first slice has not been satisfied, the operations can include performing a tradeoff logic operation to determine the final threshold value. subsequently, the operations can include determining, using the candidate machine-learned model, whether input data is authentic based on the final threshold value.
Inventor(s): Jibing Wang of San Jose CA (US) for google llc, Erik Stauffer of Sunnyvale CA (US) for google llc
IPC Code(s): G06N3/08, G06N3/04, H04L12/18
Abstract: techniques and apparatuses are described for machine-learning architectures for broadcast and multicast communications. a network entity processes broadcast or multicast communications using a deep neural network (dnn) to direct the one or more broadcast or multicast communications to a targeted group of user equipments (ues) using the wireless communication system. the network entity receives feedback from at least one user equipment (ue) of the targeted group of ues. the network entity determines a modification to the dnn based on the feedback. the network entity transmits an indication of the modification to the targeted group of ues. the network entity updates the dnn with the modification to form a modified dnn. the network entity processes the broadcast or multicast communications using the modified dnn to direct the broadcast or multicast communications to the targeted group of ues using the wireless communication system.
Inventor(s): Krishna Pragash Srinivasan of Union City CA (US) for google llc, Michael Bendersky of Cupertino CA (US) for google llc, Anupam Samanta of Mountain View CA (US) for google llc, Lingrui Liao of Foster City CA (US) for google llc, Luca Bertelli of Redwood City CA (US) for google llc, Ming-Wei Chang of Redmond WA (US) for google llc, Iftekhar Naim of San Jose CA (US) for google llc, Siddhartha Brahma of San Jose CA (US) for google llc, Siamak Shakeri of New York NY (US) for google llc, Hongkun Yu of Redwood City CA (US) for google llc, John Nham of Fremont CA (US) for google llc, Karthik Raman of Sunnyvale CA (US) for google llc, Raphael Dominik Hoffmann of Los Altos CA (US) for google llc
IPC Code(s): G06N3/0895, G06F16/903, G06F16/93, G06N3/0455
Abstract: provided are computing systems, methods, and platforms that train query processing models, such as large language models, to perform query intent classification tasks by using retrieval augmentation and multi-stage distillation. unlabeled training examples of queries may be obtained, and a set of the training examples may be augmented with additional feature annotations to generate augmented training examples. a first query processing model may annotate the retrieval augmented queries to generate inferred labels for the augmented training examples. a second query processing model may be trained on the inferred labels, distilling the query processing model that was trained with retrieval augmentation into a non-retrieval augmented query processing model. the second query processing model may annotate the entire set of unlabeled training examples. another stage of distillation may train a third query processing model using the entire set of unlabeled training examples without retrieval augmentation.
20240135217.QUANTUM CIRCUITS WITH REDUCED T GATE COUNT_simplified_abstract_(google llc)
Inventor(s): Craig Gidney of Goleta CA (US) for google llc
IPC Code(s): G06N10/20, G06N10/70, H03K19/195, H03K19/20
Abstract: methods, systems and apparatus for producing quantum circuits with low t gate counts. in one aspect, a method for performing a temporary logical and operation on two control qubits includes the actions of obtaining an ancilla qubit in an a-state; computing a logical-and of the two control qubits and storing the computed logical-and in the state of the ancilla qubit, comprising replacing the a-state of the ancilla qubit with the logical-and of the two control qubits; maintaining the ancilla qubit storing the logical-and of the two controls until a first condition is satisfied; and erasing the ancilla qubit when the first condition is satisfied.
Inventor(s): Harikrishna Narasimhan of Sunnyvale CA (US) for google llc, Wittawat Jitkrittum of Atlanta GA (US) for google llc, Aditya Krishna Menon of New York NY (US) for google llc, Ankit Singh Rawat of New York NY (US) for google llc, Sanjiv Kumar of Jericho NY (US) for google llc
IPC Code(s): G06N20/00, G06N5/04
Abstract: methods, systems, and apparatus, including computer programs encoded on computer storage media, for post-hoc deferral for classification tasks. in particular, a system can perform either post-hoc threshold correction or post-hoc rejector training to account for the cost of deferring model inputs to an expert system for classification.
Inventor(s): Omer Livne-Simha of Mountain View CA (US) for google llc, Gil Disatnik of Kibbutz Yakum (IL) for google llc
IPC Code(s): G06Q10/0631, G06Q30/018, H04W4/02, H04W4/024, H04W4/40
Abstract: to determine whether two or more devices are at the same location, a first client device receives a first short-range communication from a second client device identifying the second client device. the first client device also receives a second short-range communication from the second client device identifying the second client device. the first client device provides the first and second short-range communications to a server device that analyzes the first and second short-range communications to verify that the first and second client devices are the same location. more specifically, the server device determines a likelihood that the first and second client devices are the same location based on the first and second short-range communications. then the server devices determines a risk of fraud based on the likelihood.
20240135492.IMAGE SUPER-RESOLUTION NEURAL NETWORKS_simplified_abstract_(google llc)
Inventor(s): Cristina Nader Vasconcelos of Montreal (CA) for google llc, Ahmet Cengiz Oztireli of Zurich (CH) for google llc, Andrea Tagliasacchi of Toronto (CA) for google llc, Kevin Jordan Swersky of Toronto (CA) for google llc, Mark Jeffrey Matthews of Los Angeles CA (US) for google llc, Milad Olia Hashemi of San Francisco CA (US) for google llc
IPC Code(s): G06T3/40, G06T5/20, G06V10/771
Abstract: methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing an input image using a super-resolution neural network to generate an up-sampled image that is a higher resolution version of the input image. in one aspect, a method comprises: processing the input image using an encoder subnetwork of the super-resolution neural network to generate a feature map; generating an updated feature map, comprising, for each spatial position in the updated feature map: applying a convolutional filter to the feature map to generate a plurality of features corresponding to the spatial position in the updated feature map, wherein the convolutional filter is parametrized by a set of convolutional filter parameters that are generated by processing data representing the spatial position using a hyper neural network; and processing the updated feature map using a projection subnetwork of the super-resolution neural network to generate the up-sampled image.
Inventor(s): Ignacio Garcia Dorado of Berlin (DE) for google llc, Charles Goran of Lafayette CO (US) for google llc, Jordi Serrano Berbel of Berlin (DE) for google llc, Luke Barrington of Berlin (DE) for google llc, Bilawal Singh Sidhu of Austin TX (US) for google llc, Thomas Windheuser of Zürich (CH) for google llc, Thomas Robert Escobar of Boulder (CO) for google llc, Jan Stria of Berlin (DE) for google llc
IPC Code(s): G06T19/00, G06T7/11
Abstract: generating a location-specific three-dimensional model in response to a location query can provide users with a better understanding of a location through providing better interactivity, better perspective, and better understanding of dimensionality. generation of the models can be enabled by leveraging a three-dimensional asset database and segmentation methods. the location-specific models can provide further utility by further including situation specific simulated effects, such as simulated weather or traffic.
Inventor(s): Scott Davies of Santa Monica CA (US) for google llc, Justin Lewis of South San Francisco CA (US) for google llc
IPC Code(s): G06V20/40, G06F3/0481, G06F16/95
Abstract: a method for adaptive presentation of media content includes receiving a first version of a video content item, causing the first version of the video content item to be presented within a video viewport having first dimensions, receiving user input indicating a user request to change first dimensions of the video viewport to second dimensions, determining an area of interest in the video content item, wherein the area of interest is included in a frame of the video content item, generating, based on the second dimensions of the video viewport and the area of interest in a frame of the video content item, a second version of the video content item for presentation on a user device, and causing the second version of the video content item to be presented within the video viewport having the second dimensions on the user device, wherein the area of interest is placed differently within the frame in the second version of the video content item relative to the first version of the video content item.
Inventor(s): Jessica Lee of Brooklyn NY (US) for google llc, David Trotter Oleson of Ruschlikon (CH) for google llc, Fabian Roth of Zurich (CH) for google llc, Nils Grimsmo of Wollerau (CH) for google llc
IPC Code(s): G09B7/04, G06F3/04845, G06F40/205, G06T11/60, G06V10/94, G06V20/70, G06V30/12, G06V30/19
Abstract: systems and methods for augmented-reality tutoring can utilize optical character recognition, natural language processing, and/or augmented-reality rendering for providing real-time notifications for completing a determined task. the systems and methods can include utilizing one or more machine-learned models trained for quantitative reasoning and can include providing a plurality of different user interface elements at different times.
Inventor(s): Chien-Hui Wen of Cupertino CA (US) for google llc, Hsin-Yu Chen of Taoyuan City 300 (TW) for google llc
IPC Code(s): G09G3/20
Abstract: a method involves measuring, for an optical property of the display panel for an input gray level and at a first refresh rate, first and second values at respective first and second ambient brightness levels. the method also involves determining a compensation factor for the input gray level at the first refresh rate. the method further involves determining a modified gamma value at a second refresh rate, wherein the modified gamma value reduces a perceived optical defect of the display panel when operating at the second refresh rate by maintaining a consistent delta difference in values for the optical property between the first and second refresh rates at different ambient brightness levels. the method additionally involves storing the modified gamma value, where the device is configured to adjust input display data using the modified gamma value when transitioning from the first refresh rate to the second refresh rate.
Inventor(s): Kenneth Mixter of Los Altos Hills CA (US) for google llc, Daniel Colish of Portland OR (US) for google llc, Tuan Nguyen of San Jose CA (US) for google llc
IPC Code(s): G10L13/00, G06F3/16, G10L15/22, G10L15/26, H04L12/28, H04L51/224, H04L67/55
Abstract: a method for proactive notifications in a voice interface device includes: receiving a first user voice request for an action with an future performance time; assigning the first user voice request to a voice assistant service for performance; subsequent to the receiving, receiving a second user voice request and in response to the second user voice request initiating a conversation with the user; and during the conversation: receiving a notification from the voice assistant service of performance of the action; triggering a first audible announcement to the user to indicate a transition from the conversation and interrupting the conversation; triggering a second audible announcement to the user to indicate performance of the action; and triggering a third audible announcement to the user to indicate a transition back to the conversation and rejoining the conversation.
Inventor(s): Nobuyuki Morioka of Mountain View CA (US) for google llc, Byungha Chun of Tokyo (JP) for google llc, Nanxin Chen of Mountain View CA (US) for google llc, Yu Zhang of Mountain View CA (US) for google llc, Yifan Ding of Mountain View CA (US) for google llc
IPC Code(s): G10L13/027
Abstract: a method for residual adapters for few-shot text-to-speech speaker adaptation includes obtaining a text-to-speech (tts) model configured to convert text into representations of synthetic speech, the tts model pre-trained on an initial training data set. the method further includes augmenting the tts model with a stack of residual adapters. the method includes receiving an adaption training data set including one or more spoken utterances spoken by a target speaker, each spoken utterance in the adaptation training data set paired with corresponding input text associated with a transcription of the spoken utterance. the method also includes adapting, using the adaption training data set, the tts model augmented with the stack of residual adapters to learn how to synthesize speech in a voice of the target speaker by optimizing the stack of residual adapters while parameters of the tts model are frozen.
Inventor(s): Tien-Ju Yang of Mountain View CA (US) for google llc, You-Chi Cheng of Mountain View CA (US) for google llc, Shankar Kumar of New York NY (US) for google llc, Jared Lichtarge of Mountain View CA (US) for google llc, Ehsan Amid of Mountain View CA (US) for google llc, Yuxin Ding of Mountain View CA (US) for google llc, Rajiv Mathews of Sunnyvale CA (US) for google llc, Mingqing Chen of Saratoga CA (US) for google llc
IPC Code(s): G10L15/06, G10L15/197, G10L15/30
Abstract: a method includes receiving distillation data including a plurality of out-of-domain training utterances. for each particular out-of-domain training utterance of the distillation data, the method includes generating a corresponding augmented out-of-domain training utterance, and generating, using a teacher asr model trained on training data corresponding to a target domain, a pseudo-label corresponding to the corresponding augmented out-of-domain training utterance. the method also includes distilling a student asr model from the teacher asr model by training the student asr model using the corresponding augmented out-of-domain training utterances paired with the corresponding pseudo-labels generated by the teacher asr model.
Inventor(s): Chao Zhang of Mountain View CA (US) for google llc, Bo Li of Santa Clara CA (US) for google llc, Tara N. Sainath of Jersey City NJ (US) for google llc, Trevor Strohman of Mountain View CA (US) for google llc, Shuo-yiin Chang of Sunnyvale CA (US) for google llc
IPC Code(s): G10L15/197, G10L15/00, G10L15/02
Abstract: a method includes receiving a sequence of acoustic frames as input to a multilingual automated speech recognition (asr) model configured to recognize speech in a plurality of different supported languages and generating, by an audio encoder of the multilingual asr, a higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. the method also includes generating, by a language identification (lid) predictor of the multilingual asr, a language prediction representation for a corresponding higher order feature representation. the method also includes generating, by a decoder of the multilingual asr, a probability distribution over possible speech recognition results based on the corresponding higher order feature representation, a sequence of non-blank symbols, and a corresponding language prediction representation. the decoder includes monolingual output layer having a plurality of output nodes each sharing a plurality of language-specific wordpiece models.
Inventor(s): Johnny Chen of Sunnyvale CA (US) for google llc, Thomas L. Dean of Los Altos Hills CA (US) for google llc, Qiangfeng Peter Lau of Mountain View CA (US) for google llc, Sudeep Gandhe of Sunnyvale CA (US) for google llc, Gabriel Schine of Los Banos CA (US) for google llc
IPC Code(s): G10L15/22, G06F16/332, G06F21/62, G10L13/00, G10L13/033, G10L13/08, G10L15/18, G10L15/26, G10L15/30, G10L17/22, H04L67/104
Abstract: methods, systems, and apparatus, including computer programs encoded on computer storage media, for handing off a user conversation between computer-implemented agents. one of the methods includes receiving, by a computer-implemented agent specific to a user device, a digital representation of speech encoding an utterance, determining, by the computer-implemented agent, that the utterance specifies a requirement to establish a communication with another computer-implemented agent, and establishing, by the computer-implemented agent, a communication between the other computer-implemented agent and the user device.
20240135931.TRANSCRIPTION BASED ON SPEECH AND VISUAL INPUT_simplified_abstract_(google llc)
Inventor(s): Xavier Benavides Palos of Beverly Hills CA (US) for google llc
IPC Code(s): G10L15/26, G06V20/40, G10L15/18
Abstract: a method can include receiving audio input of speech, receiving visual input while receiving the audio input, generating a semantic description based on the visual input, and presenting a transcription of the speech based on the audio input and the semantic description.
Inventor(s): Guanlong Zhao of Long Island City NY (US) for google llc, Quan Wang of Hoboken NJ (US) for google llc, Han Lu of Redmond WA (US) for google llc, Yiling Huang of Edgewater NJ (US) for google llc, Jason Pelecanos of Mountain View CA (US) for google llc
IPC Code(s): G10L17/06, G10L17/02, G10L17/04
Abstract: a method includes obtaining a multi-utterance training sample that includes audio data characterizing utterances spoken by two or more different speakers and obtaining ground-truth speaker change intervals indicating time intervals in the audio data where speaker changes among the two or more different speakers occur. the method also includes processing the audio data to generate a sequence of predicted speaker change tokens using a sequence transduction model. for each corresponding predicted speaker change token, the method includes labeling the corresponding predicted speaker change token as correct when the predicted speaker change token overlaps with one of the ground-truth speaker change intervals. the method also includes determining a precision metric of the sequence transduction model based on a number of the predicted speaker change tokens labeled as correct and a total number of the predicted speaker change tokens in the sequence of predicted speaker change tokens.
Inventor(s): Chaitanya GHARPURE of Santa Clara CA (US) for google llc, Evan FISHER of San Francisco CA (US) for google llc, Eric LIU of Redwood City CA (US) for google llc, Peng YANG of San Jose CA (US) for google llc, Emily HOU of Mountain View CA (US) for google llc, Victoria FANG of Mountain View CA (US) for google llc
IPC Code(s): G10L25/87, G06F3/0483, G10L15/02, G10L15/04, G10L15/10, G10L15/26
Abstract: the disclosure provides technology for enhancing the ability of a computing device to detect when a user has discontinued reading a text source. an example method includes receiving audio data comprising a spoken word associated with a text source, comparing the audio data with data of the text source, determining, based on the comparing, whether a segment of the audio data corresponds to a location of the text source, and responsive to determining that the segment of the audio data does not correspond to a location of the text source, transmitting a signal indicating that a user has discontinued reading the text source, the signal causing to cease the comparing of the audio data with the data of the text source.
20240136857.WIRELESS CHARGING SYSTEM WITH RECEIVER CONTROL_simplified_abstract_(google llc)
Inventor(s): Li Wang of Mountain View CA (US) for google llc, Liang Jia of Palo Alto CA (US) for google llc
IPC Code(s): H02J50/10, H02J50/40, H02M7/219
Abstract: an example device () includes a wireless charging receive coil () configured to transduce, into an alternating current—ac—power signal, a magnetic field generated by a wireless charging transmit coil () of an external device (); an active rectifier () configured to convert the ac signal into a direct current—dc—power signal; and circuitry () configured to: obtain a target level of the dc power signal; and control the active rectifier () to output the dc power signal with the target level.
Inventor(s): Jibing Wang of San Jose CA (US) for google llc, Erik Richard Stauffer of Sunnyvale CA (US) for google llc
IPC Code(s): H04B7/024, H04B7/06
Abstract: techniques described herein describe aspects of signal adjustments in user equipment-coordination set, uecs, joint transmissions. a base station analyzes a first joint transmission from multiple user equipments, ues, participating in a uecs, where the multiple ues include a coordinating ue of the uecs and at least one non-coordinating ue participating in the uecs. the base station determines that the first joint transmission fails to meet a performance metric and directs the multiple ues participating in the uecs to add signal adjustments to a second joint transmission.
Inventor(s): Abhishek Agarwal of Santa Clara CA (US) for google llc, Ye Tang of Palo Alto CA (US) for google llc, Prashant R. Chandra of San Jose CA (US) for google llc, Simon Luigi Sabato of Saratoga CA (US) for google llc, Hema Hariharan of Cupertino CA (US) for google llc
IPC Code(s): H04J3/06
Abstract: aspects of the disclosure are directed to supporting time synchronization across a datacenter network with greater accuracy. the time synchronization includes both software based and hardware based time synchronization mechanisms to provide more precise time synchronization across various nodes in the datacenter network. the software based mechanism can provide the initial coarse time synchronization while the hardware based mechanism can provide the subsequent finer time synchronization.
20240137308.Virtual Channel Balancing In Ring-Based Topologies_simplified_abstract_(google llc)
Inventor(s): Brian Patrick Towles of Chapel Hill NC (US) for google llc, Hojat Parta of Los Angeles CA (US) for google llc
IPC Code(s): H04L45/00, H04L49/102
Abstract: systems and method for routing data packets in ring network. a data packet being transmitted to a destination node may be received by a first structure at a first node. the first node may determine a number of hops the data packet will traverse as it is transmitted from the first node to the destination node and compare the determined number of hops to a threshold hop value to determine whether the number of hops is equal to or less than the threshold hop value. if the number of hops is greater than the threshold, the data packet may be transmitted to a dimension queuing structure for a first virtual channel within a second node, otherwise, the data packet may be transmitted to a dimension queuing structure for a second virtual channel or a turn queuing structure within the second node.
Inventor(s): Ching Yin Derek Pang of San Jose CA (US) for google llc, Kyrah Felder of Kennesaw GA (US) for google llc, Akshay Gadde of Fremont CA (US) for google llc, Paul Wilkins of Cambridge (GB) for google llc, Cheng Chen of Milpitas CA (US) for google llc, Yao-Chung Lin of Sunnyvale CA (US) for google llc
IPC Code(s): H04L65/70, G06N20/00, H04L65/61, H04L65/80, H04N21/25
Abstract: a media item to be provided to users of a platform is identified. the media item is associated with a media class of one or more media classes. an indication of the media item is provided as input to a machine learning model trained based on historical encoding data to predict, for a given media item, a set of encoder parameter settings that satisfy a performance criterion in view of a respective media class of the given media item. the historical encoding data includes a prior set of encoder parameter settings that satisfied the performance criterion with respect to a prior media item associated with the respective class. encoder parameter settings that satisfy the performance criterion in view of the media class is determined based on an output of the model. the media item is caused to be encoded using the determined encoder parameter settings.
20240137423.PROACTIVE ENVIRONMENT-BASED CHAT INFORMATION SYSTEM_simplified_abstract_(google llc)
Inventor(s): Ilya Gennadyevich Gelfenbeyn of Sunnyvale CA (US) for google llc, Artem Goncharuk of Mountain View CA (US) for google llc, Ilya Andreevich Platonov of Berdsk (RU) for google llc, Pavel Aleksandrovich Sirotin of Sunnyvale CA (US) for google llc, Olga Aleksandrovna Gelfenbeyn of Yurga (RU) for google llc
IPC Code(s): H04L67/55, H04L51/02
Abstract: disclosed is the technology for computer-based “daily brief” service, which includes methods and corresponding systems for proactively providing push notifications for users of chat information systems. the push notifications are dynamically generated and presented to the user based on identification of one or more triggering events, which may include predetermined time/date, current geographical location, activity of peers and friends in social media associated with the user, scheduled events, appointments, meetings, emails, instant messages, and many more. the described technology improves the interaction interface between the user and chat information system.
Inventor(s): Stéphane Hervé Loïc Hulaud of Stockholm (SE) for google llc
IPC Code(s): H04N7/15, G06F3/01, G06T7/70, G06V10/141, G06V10/26, G06V10/70
Abstract: methods, systems, and apparatus are described for immersive videoconferencing teleconferencing streams from multiple endpoints within shared scene environment. the method includes receiving a plurality of streams for presentation at a teleconference, wherein each of the plurality of streams represents a participant of a respective plurality of participants of the teleconference. the method includes, determining scene data descriptive of a scene environment, the scene data comprising at least one of lighting characteristics, acoustic characteristics, or perspective characteristics of the scene environment. the method includes, for each of the plurality of participants of the teleconference, determining a position of the participant within the scene environment and, based at least in part on the scene data and the position of the participant within the scene environment, modifying the stream that represents the participant.
Inventor(s): Damien Kelly of San Francisco CA (US) for google llc, Bartlomiej Wronski of Santa Monica CA (US) for google llc
IPC Code(s): H04N19/65, H04N19/60
Abstract: a video item that is subject to one or more motion stabilization transformations applied to the video item is identified. the one or more motion stabilization transformations pertain to a motion between video frames of a video sequence associated with the video item. one or more operations to modify the one or more motion stabilization transformations applied to the video item are determined. the one or more operations to modify the motion stabilization transformations are performed. the video item is provided for playback via a client device. playback of the video item depicts at least a portion of the motion between the video frames of the video sequence.
20240137674.Optical Link Diagnostic System_simplified_abstract_(google llc)
Inventor(s): Jill Berger of Saratoga CA (US) for google llc, Kevin Yasumura of Lafayette CA (US) for google llc, Xiang Zhou of Sunnyvale CA (US) for google llc, Pedram Z. Dashti of San Jose CA (US) for google llc, Ryohei Urata of San Carlos CA (US) for google llc
IPC Code(s): H04Q11/00, G02B6/35, G02B26/08, G02B27/30, H04B10/071, H04B10/079
Abstract: an optical links diagnostic system (lds) and its operation within an optical circuit switch (ocs) for measurement and diagnosis of fiber-optic network fiber performance and quality is disclosed. the lds can contain two photodetectors, a laser source, and be coupled to an ocs. optical circulators can further be linked to the ocs. the lds can be used both as an optical time domain reflectometer (otdr) or as an optical return loss (orl) meter and can automate the diagnosis of the fiber optical network fiber insertion loss and return loss.
Inventor(s): Mahesh Devdatta Telang of Mountain View CA (US) for google llc, Qin Zhang of Mountain View CA (US) for google llc, Shivank Nayak of Milpitas CA (US) for google llc, Rukun Mao of Santa Clara CA (US) for google llc
IPC Code(s): H04W48/02, H04W8/24, H04W60/00
Abstract: the present disclosure describes apparatuses and techniques of adaptive plmn management for varying network conditions. in some aspects, a plmn manager of a user equipment (ue) searches, as part of a registration procedure, a data repository of plmn information that includes a plmn blocked from registration for a duration of time and determines that the plmn is blocked due to a signal-related condition with a cell of the plmn (e.g., weak signal). the plmn manager then determines, during the duration of time, that the signal-related condition between the ue and the cell has improved. in response to the improvement, the plmn manager unblocks the plmn before expiration of the duration of time. by so doing, the ue may reattempt registration with the unblocked plmn at least once before the duration of time expires, which may allow the ue to register more quickly and reestablish network services.
20240138004.MANAGING A CELL GROUP IN DUAL CONNECTIVITY_simplified_abstract_(google llc)
Inventor(s): Chih-Hsiang Wu of Taoyuan City (TW) for google llc
IPC Code(s): H04W76/15, H04W76/30
Abstract: a network node of a radio access network (ran), communicating with a user equipment (ue) in dual connectivity (dc) with a master node (mn) and a secondary node (sn), can implement a method for managing deactivation of a secondary cell group (scg). the method includes detecting () that a condition for deactivating the scg is satisfied. the method further includes determining () whether the ue supports deactivating the scg, and causing () the sn to deactivate or to release the scg at the sn based on the determining.
- GOOGLE LLC
- B25J9/16
- G05B13/02
- G05B19/042
- G06N3/008
- G06N3/045
- G06N3/08
- Google llc
- G01C21/36
- G01C21/28
- G06F18/214
- G06T7/80
- G06V20/10
- H04N23/698
- G02B27/01
- G02B6/34
- G03F7/00
- G03F7/40
- G06F3/01
- G01V1/00
- G06V40/20
- G06F9/50
- G06F9/455
- G06F16/242
- G06N20/00
- G06F16/332
- G06F16/00
- G06F16/33
- G06F40/30
- G06Q10/10
- G10L15/22
- G06F16/955
- G06F16/23
- G06F21/56
- G06F21/52
- G06N3/04
- G06F21/57
- G06F21/44
- G06F21/64
- G06F40/58
- G10L15/02
- G10L15/30
- G10L19/032
- G06K7/14
- G06K19/06
- G06V10/20
- G06N3/048
- H04L12/18
- G06N3/0895
- G06F16/903
- G06F16/93
- G06N3/0455
- G06N10/20
- G06N10/70
- H03K19/195
- H03K19/20
- G06N5/04
- G06Q10/0631
- G06Q30/018
- H04W4/02
- H04W4/024
- H04W4/40
- G06T3/40
- G06T5/20
- G06V10/771
- G06T19/00
- G06T7/11
- G06V20/40
- G06F3/0481
- G06F16/95
- G09B7/04
- G06F3/04845
- G06F40/205
- G06T11/60
- G06V10/94
- G06V20/70
- G06V30/12
- G06V30/19
- G09G3/20
- G10L13/00
- G06F3/16
- G10L15/26
- H04L12/28
- H04L51/224
- H04L67/55
- G10L13/027
- G10L15/06
- G10L15/197
- G10L15/00
- G06F21/62
- G10L13/033
- G10L13/08
- G10L15/18
- G10L17/22
- H04L67/104
- G10L17/06
- G10L17/02
- G10L17/04
- G10L25/87
- G06F3/0483
- G10L15/04
- G10L15/10
- H02J50/10
- H02J50/40
- H02M7/219
- H04B7/024
- H04B7/06
- H04J3/06
- H04L45/00
- H04L49/102
- H04L65/70
- H04L65/61
- H04L65/80
- H04N21/25
- H04L51/02
- H04N7/15
- G06T7/70
- G06V10/141
- G06V10/26
- G06V10/70
- H04N19/65
- H04N19/60
- H04Q11/00
- G02B6/35
- G02B26/08
- G02B27/30
- H04B10/071
- H04B10/079
- H04W48/02
- H04W8/24
- H04W60/00
- H04W76/15
- H04W76/30
