Patents by Inventor Ehsan Variani
Ehsan Variani has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).
-
Patent number: 11922322Abstract: Aspects of the present disclosure enable humanly-specified relationships to contribute to a mapping that enables compression of the output structure of a machine-learned model. An exponential model such as a maximum entropy model can leverage a machine-learned embedding and the mapping to produce a classification output. In such fashion, the feature discovery capabilities of machine-learned models (e.g., deep networks) can be synergistically combined with relationships developed based on human understanding of the structural nature of the problem to be solved, thereby enabling compression of model output structures without significant loss of accuracy. These compressed models provide improved applicability to “on device” or other resource-constrained scenarios.Type: GrantFiled: January 30, 2023Date of Patent: March 5, 2024Assignee: GOOGLE LLCInventors: Mitchel Weintraub, Ananda Theertha Suresh, Ehsan Variani
-
Publication number: 20230343328Abstract: An ASR model includes a first encoder configured to receive a sequence of acoustic frames and generate a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The ASR model also includes a second encoder configured to receive the first higher order feature representation generated by the first encoder at each of the plurality of output steps and generate a second higher order feature representation for a corresponding first higher order feature frame. The ASR model also includes a decoder configured to receive the second higher order feature representation generated by the second encoder at each of the plurality of output steps and generate a first probability distribution over possible speech recognition hypothesis. The ASR model also includes a language model configured to receive the first probability distribution over possible speech hypothesis and generate a rescored probability distribution.Type: ApplicationFiled: June 16, 2023Publication date: October 26, 2023Applicant: Google LLCInventors: Tara Sainath, Arun Narayanan, Rami Botros, Yanzhang He, Ehsan Variani, Cyril Allauzen, David Rybach, Ruoming Pang, Trevor Strohman
-
Patent number: 11783849Abstract: This specification describes computer-implemented methods and systems. One method includes receiving, by a neural network of a speech recognition system, first data representing a first raw audio signal and second data representing a second raw audio signal. The first raw audio signal and the second raw audio signal describe audio occurring at a same period of time. The method further includes generating, by a spatial filtering layer of the neural network, a spatial filtered output using the first data and the second data, and generating, by a spectral filtering layer of the neural network, a spectral filtered output using the spatial filtered output. Generating the spectral filtered output comprises processing frequency-domain data representing the spatial filtered output. The method still further includes processing, by one or more additional layers of the neural network, the spectral filtered output to predict sub-word units encoded in both the first raw audio signal and the second raw audio signal.Type: GrantFiled: June 8, 2021Date of Patent: October 10, 2023Assignee: Google LLCInventors: Ehsan Variani, Kevin William Wilson, Ron J. Weiss, Tara N. Sainath, Arun Narayanan
-
Publication number: 20230298570Abstract: A method includes generating, using an audio encoder, a higher-order feature representation for each acoustic frame in a sequence of acoustic frames; generating, using a decoder, based on the higher-order feature representation, a plurality of speech recognition hypotheses, each hypotheses corresponding to a candidate transcription of an utterance and having an associated first likelihood score; generating, using an external language model, for each speech recognition hypothesis, a second likelihood score; determining, using a learnable fusion module, for each speech recognition hypothesis, a set of fusion weights based on the higher-order feature representation and the speech recognition hypothesis; and generating, using the learnable fusion module, for each speech recognition hypothesis, a third likelihood score based on the first likelihood score, the second likelihood score, and the set of fusion weights, the audio encoder and decoder trained using minimum additive error rate training in the presence of tType: ApplicationFiled: March 21, 2023Publication date: September 21, 2023Applicant: Google LLCInventors: Weiran Wang, Tongzhou Chen, Tara N. Sainath, Ehsan Variani, Rohit Prakash Prabhavalkar, Ronny Huang, Bhuvana Ramabhadran, Neeraj Gaur, Sepand Mavandadi, Charles Caleb Peyser, Trevor Strohman, Yangzhang He, David Rybach
-
Patent number: 11715458Abstract: An ASR model includes a first encoder configured to receive a sequence of acoustic frames and generate a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The ASR model also includes a second encoder configured to receive the first higher order feature representation generated by the first encoder at each of the plurality of output steps and generate a second higher order feature representation for a corresponding first higher order feature frame. The ASR model also includes a decoder configured to receive the second higher order feature representation generated by the second encoder at each of the plurality of output steps and generate a first probability distribution over possible speech recognition hypothesis. The ASR model also includes a language model configured to receive the first probability distribution over possible speech hypothesis and generate a rescored probability distribution.Type: GrantFiled: May 10, 2021Date of Patent: August 1, 2023Assignee: Google LLCInventors: Tara Sainath, Arun Narayanan, Rami Botros, Yanzhang He, Ehsan Variani, Cyril Allauzen, David Rybach, Ruoming Pang, Trevor Strohman
-
Publication number: 20230186096Abstract: Aspects of the present disclosure enable humanly-specified relationships to contribute to a mapping that enables compression of the output structure of a machine-learned model. An exponential model such as a maximum entropy model can leverage a machine-learned embedding and the mapping to produce a classification output. In such fashion, the feature discovery capabilities of machine-learned models (e.g., deep networks) can be synergistically combined with relationships developed based on human understanding of the structural nature of the problem to be solved, thereby enabling compression of model output structures without significant loss of accuracy. These compressed models provide improved applicability to “on device” or other resource-constrained scenarios.Type: ApplicationFiled: January 30, 2023Publication date: June 15, 2023Inventors: Mitchel Weintraub, Ananda Theertha Suresh, Ehsan Variani
-
Patent number: 11568260Abstract: Aspects of the present disclosure enable humanly-specified relationships to contribute to a mapping that enables compression of the output structure of a machine-learned model. An exponential model such as a maximum entropy model can leverage a machine-learned embedding and the mapping to produce a classification output. In such fashion, the feature discovery capabilities of machine-learned models (e.g., deep networks) can be synergistically combined with relationships developed based on human understanding of the structural nature of the problem to be solved, thereby enabling compression of model output structures without significant loss of accuracy. These compressed models provide improved applicability to “on device” or other resource-constrained scenarios.Type: GrantFiled: October 16, 2019Date of Patent: January 31, 2023Assignee: GOOGLE LLCInventors: Mitchel Weintraub, Ananda Theertha Suresh, Ehsan Variani
-
Publication number: 20220310081Abstract: A method includes receiving a sequence of acoustic frames extracted from audio data corresponding to an utterance. During a first pass, the method includes processing the sequence of acoustic frames to generate N candidate hypotheses for the utterance. During a second pass, and for each candidate hypothesis, the method includes generating a respective un-normalized likelihood score; generating a respective external language model score; generating a standalone score that models prior statistics of the corresponding candidate hypothesis, and generating a respective overall score for the candidate hypothesis based on the un-normalized likelihood score, the external language model score, and the standalone score. The method also includes selecting the candidate hypothesis having the highest respective overall score from among the N candidate hypotheses as a final transcription of the utterance.Type: ApplicationFiled: March 22, 2022Publication date: September 29, 2022Applicant: Google LLCInventors: Neeraj Gaur, Tongzhou Chen, Ehsan Variani, Bhuvana Ramabhadran, Parisa Haghani, Pedro J. Moreno Mengibar
-
Publication number: 20220310062Abstract: An ASR model includes a first encoder configured to receive a sequence of acoustic frames and generate a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The ASR model also includes a second encoder configured to receive the first higher order feature representation generated by the first encoder at each of the plurality of output steps and generate a second higher order feature representation for a corresponding first higher order feature frame. The ASR model also includes a decoder configured to receive the second higher order feature representation generated by the second encoder at each of the plurality of output steps and generate a first probability distribution over possible speech recognition hypothesis. The ASR model also includes a language model configured to receive the first probability distribution over possible speech hypothesis and generate a rescored probability distribution.Type: ApplicationFiled: May 10, 2021Publication date: September 29, 2022Applicant: Google LLCInventors: Tara Sainath, Arun Narayanan, Rami Botros, Yangzhang He, Ehsan Variani, Cyrill Allauzen, David Rybach, Ruorning Pang, Trevor Strohman
-
Patent number: 11393457Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition using complex linear projection are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The method further includes generating frequency domain data using the audio data. The method further includes processing the frequency domain data using complex linear projection. The method further includes providing the processed frequency domain data to a neural network trained as an acoustic model. The method further includes generating a transcription for the utterance that is determined based at least on output that the neural network provides in response to receiving the processed frequency domain data.Type: GrantFiled: May 20, 2020Date of Patent: July 19, 2022Assignee: Google LLCInventors: Samuel Bengio, Mirko Visontai, Christopher Walter George Thornton, Tara N. Sainath, Ehsan Variani, Izhak Shafran, Michiel A. u. Bacchiani
-
Publication number: 20220122622Abstract: An automated speech recognition (ASR) model includes a first encoder, a second encoder, and a decoder. The first encoder receives, as input, a sequence of acoustic frames, and generates, at each of a plurality of output steps, a first higher order feature representation for a corresponding acoustic frame in the sequence of acoustic frames. The second encoder receives, as input, the first higher order feature representation generated by the first encoder at each of the plurality of output steps, and generates, at each of the plurality of output steps, a second higher order feature representation for a corresponding first higher order feature frame. The decoder receives, as input, the second higher order feature representation generated by the second encoder at each of the plurality of output steps, and generates, at each of the plurality of time steps, a first probability distribution over possible speech recognition hypotheses.Type: ApplicationFiled: April 21, 2021Publication date: April 21, 2022Applicant: Google LLCInventors: Arun Narayanan, Tara Sainath, Chung-Cheng Chiu, Ruoming Pang, Rohit Prabhavalkar, Jiahui Yu, Ehsan Variani, Trevor Strohman
-
Publication number: 20210295859Abstract: This specification describes computer-implemented methods and systems. One method includes receiving, by a neural network of a speech recognition system, first data representing a first raw audio signal and second data representing a second raw audio signal. The first raw audio signal and the second raw audio signal describe audio occurring at a same period of time. The method further includes generating, by a spatial filtering layer of the neural network, a spatial filtered output using the first data and the second data, and generating, by a spectral filtering layer of the neural network, a spectral filtered output using the spatial filtered output. Generating the spectral filtered output comprises processing frequency-domain data representing the spatial filtered output. The method still further includes processing, by one or more additional layers of the neural network, the spectral filtered output to predict sub-word units encoded in both the first raw audio signal and the second raw audio signal.Type: ApplicationFiled: June 8, 2021Publication date: September 23, 2021Applicant: Google LLCInventors: Ehsan Variani, Kevin William Wilson, Ron J. Weiss, Tara N. Sainath, Arun Narayanan
-
Patent number: 11062725Abstract: This specification describes computer-implemented methods and systems. One method includes receiving, by a neural network of a speech recognition system, first data representing a first raw audio signal and second data representing a second raw audio signal. The first raw audio signal and the second raw audio signal describe audio occurring at a same period of time. The method further includes generating, by a spatial filtering layer of the neural network, a spatial filtered output using the first data and the second data, and generating, by a spectral filtering layer of the neural network, a spectral filtered output using the spatial filtered output. Generating the spectral filtered output comprises processing frequency-domain data representing the spatial filtered output. The method still further includes processing, by one or more additional layers of the neural network, the spectral filtered output to predict sub-word units encoded in both the first raw audio signal and the second raw audio signal.Type: GrantFiled: February 19, 2019Date of Patent: July 13, 2021Assignee: Google LLCInventors: Ehsan Variani, Kevin William Wilson, Ron J. Weiss, Tara N. Sainath, Arun Narayanan
-
Publication number: 20200286468Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition using complex linear projection are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The method further includes generating frequency domain data using the audio data. The method further includes processing the frequency domain data using complex linear projection. The method further includes providing the processed frequency domain data to a neural network trained as an acoustic model. The method further includes generating a transcription for the utterance that is determined based at least on output that the neural network provides in response to receiving the processed frequency domain data.Type: ApplicationFiled: May 20, 2020Publication date: September 10, 2020Applicant: Google LLCInventors: Samuel Bengio, Mirko Visontai, Christopher Walter George Thornton, Tara N. Sainath, Ehsan Variani, Izhak Shafran, Michiel A.u. Bacchiani
-
Patent number: 10714078Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition using complex linear projection are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The method further includes generating frequency domain data using the audio data. The method further includes processing the frequency domain data using complex linear projection. The method further includes providing the processed frequency domain data to a neural network trained as an acoustic model. The method further includes generating a transcription for the utterance that is determined based at least on output that the neural network provides in response to receiving the processed frequency domain data.Type: GrantFiled: October 26, 2018Date of Patent: July 14, 2020Assignee: Google LLCInventors: Samuel Bengio, Mirkó Visontai, Christopher Walter George Thornton, Michiel A. U. Bacchiani, Tara N. Sainath, Ehsan Variani, Izhak Shafran
-
Publication number: 20200134466Abstract: Aspects of the present disclosure enable humanly-specified relationships to contribute to a mapping that enables compression of the output structure of a machine-learned model. An exponential model such as a maximum entropy model can leverage a machine-learned embedding and the mapping to produce a classification output. In such fashion, the feature discovery capabilities of machine-learned models (e.g., deep networks) can be synergistically combined with relationships developed based on human understanding of the structural nature of the problem to be solved, thereby enabling compression of model output structures without significant loss of accuracy. These compressed models provide improved applicability to “on device” or other resource-constrained scenarios.Type: ApplicationFiled: October 16, 2019Publication date: April 30, 2020Inventors: Mitchel Weintraub, Ananda Theertha Suresh, Ehsan Variani
-
Publication number: 20190259409Abstract: This specification describes computer-implemented methods and systems. One method includes receiving, by a neural network of a speech recognition system, first data representing a first raw audio signal and second data representing a second raw audio signal. The first raw audio signal and the second raw audio signal describe audio occurring at a same period of time. The method further includes generating, by a spatial filtering layer of the neural network, a spatial filtered output using the first data and the second data, and generating, by a spectral filtering layer of the neural network, a spectral filtered output using the spatial filtered output. Generating the spectral filtered output comprises processing frequency-domain data representing the spatial filtered output. The method still further includes processing, by one or more additional layers of the neural network, the spectral filtered output to predict sub-word units encoded in both the first raw audio signal and the second raw audio signal.Type: ApplicationFiled: February 19, 2019Publication date: August 22, 2019Inventors: Ehsan Variani, Kevin William Wilson, Ron J. Weiss, Tara N. Sainath, Arun Narayanan
-
Publication number: 20190115013Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition using complex linear projection are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The method further includes generating frequency domain data using the audio data. The method further includes processing the frequency domain data using complex linear projection. The method further includes providing the processed frequency domain data to a neural network trained as an acoustic model. The method further includes generating a transcription for the utterance that is determined based at least on output that the neural network provides in response to receiving the processed frequency domain data.Type: ApplicationFiled: October 26, 2018Publication date: April 18, 2019Inventors: Samuel Bengio, Mirko Visontai, Christopher Walter George Thornton, Michiel A.U. Bacchiani, Tara N. Sainath, Ehsan Variani, Izhak Shafran
-
Patent number: 10224058Abstract: This specification describes computer-implemented methods and systems. One method includes receiving, by a neural network of a speech recognition system, first data representing a first raw audio signal and second data representing a second raw audio signal. The first raw audio signal and the second raw audio signal describe audio occurring at a same period of time. The method further includes generating, by a spatial filtering layer of the neural network, a spatial filtered output using the first data and the second data, and generating, by a spectral filtering layer of the neural network, a spectral filtered output using the spatial filtered output. Generating the spectral filtered output comprises processing frequency-domain data representing the spatial filtered output. The method still further includes processing, by one or more additional layers of the neural network, the spectral filtered output to predict sub-word units encoded in both the first raw audio signal and the second raw audio signal.Type: GrantFiled: November 14, 2016Date of Patent: March 5, 2019Assignee: Google LLCInventors: Ehsan Variani, Kevin William Wilson, Ron J. Weiss, Tara N. Sainath, Arun Narayanan
-
Patent number: 10140980Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for speech recognition using complex linear projection are disclosed. In one aspect, a method includes the actions of receiving audio data corresponding to an utterance. The method further includes generating frequency domain data using the audio data. The method further includes processing the frequency domain data using complex linear projection. The method further includes providing the processed frequency domain data to a neural network trained as an acoustic model. The method further includes generating a transcription for the utterance that is determined based at least on output that the neural network provides in response to receiving the processed frequency domain data.Type: GrantFiled: December 21, 2016Date of Patent: November 27, 2018Assignee: Google LCCInventors: Samuel Bengio, Mirko Visontai, Christopher Walter George Thornton, Michiel A. U. Bacchiani, Tara N. Sainath, Ehsan Variani, Izhak Shafran