Patents by Inventor Steven J. Rennie
Steven J. Rennie 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).
-
Publication number: 20190147355Abstract: Machine logic for: (i) selecting a sampled word for use as a next word in a text stream; (ii) determining, by an algorithm, an expected future reward value for the sampled word using a test policy including a training policy and a test-time inference procedure; and (iii) normalizing a set of expected future reward estimate(s) using the expected future reward value for the sampled word.Type: ApplicationFiled: November 14, 2017Publication date: May 16, 2019Inventors: Steven J. Rennie, Etienne Marcheret, Youssef Mroueh, Vaibhava Goel, Jarret Ross, Pierre L. Dognin
-
Patent number: 10217456Abstract: A method and system for generating training data for a target domain using speech data of a source domain. The training data generation method including: reading out a Gaussian mixture model (GMM) of a target domain trained with a clean speech data set of the target domain; mapping, by referring to the GMM of the target domain, a set of source domain speech data received as an input to the set of target domain speech data on a basis of a channel characteristic of the target domain speech data; and adding a noise of the target domain to the mapped set of source domain speech data to output a set of pseudo target domain speech data.Type: GrantFiled: April 14, 2014Date of Patent: February 26, 2019Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Osamu Ichikawa, Steven J Rennie
-
Patent number: 9741341Abstract: A speech processing method and arrangement are described. A dynamic noise adaptation (DNA) model characterizes a speech input reflecting effects of background noise. A null noise DNA model characterizes the speech input based on reflecting a null noise mismatch condition. A DNA interaction model performs Bayesian model selection and re-weighting of the DNA model and the null noise DNA model to realize a modified DNA model characterizing the speech input for automatic speech recognition and compensating for noise to a varying degree depending on relative probabilities of the DNA model and the null noise DNA model.Type: GrantFiled: January 20, 2015Date of Patent: August 22, 2017Assignee: Nuance Communications, Inc.Inventors: Steven J. Rennie, Pierre Dognin, Petr Fousek
-
Patent number: 9251784Abstract: A method and apparatus are provided for training a transformation matrix of a feature vector for an acoustic model. The method includes training the transformation matrix of the feature vector. The transformation matrix maximizes an objective function having a regularization term. The method further includes transforming the feature vector using the transformation matrix of the feature vector, and updating the acoustic model stored in a memory device using the transformed feature vector.Type: GrantFiled: October 23, 2013Date of Patent: February 2, 2016Assignee: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Takashi Fukuda, Vaibhava Goel, Steven J. Rennie
-
Publication number: 20150199964Abstract: A speech processing method and arrangement are described. A dynamic noise adaptation (DNA) model characterizes a speech input reflecting effects of background noise. A null noise DNA model characterizes the speech input based on reflecting a null noise mismatch condition. A DNA interaction model performs Bayesian model selection and re-weighting of the DNA model and the null noise DNA model to realize a modified DNA model characterizing the speech input for automatic speech recognition and compensating for noise to a varying degree depending on relative probabilities of the DNA model and the null noise DNA model.Type: ApplicationFiled: January 20, 2015Publication date: July 16, 2015Inventors: Steven J. Rennie, Pierre Dognin, Petr Fousek
-
Publication number: 20150112669Abstract: A method and apparatus are provided for training a transformation matrix of a feature vector for an acoustic model. The method includes training the transformation matrix of the feature vector. The transformation matrix maximizes an objective function having a regularization term. The method further includes transforming the feature vector using the transformation matrix of the feature vector, and updating the acoustic model stored in a memory device using the transformed feature vector.Type: ApplicationFiled: October 23, 2013Publication date: April 23, 2015Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Takashi Fukuda, Vaibhava Goel, Steven J. Rennie
-
Patent number: 8972258Abstract: Techniques disclosed herein include using a Maximum A Posteriori (MAP) adaptation process that imposes sparseness constraints to generate acoustic parameter adaptation data for specific users based on a relatively small set of training data. The resulting acoustic parameter adaptation data identifies changes for a relatively small fraction of acoustic parameters from a baseline acoustic speech model instead of changes to all acoustic parameters. This results in user-specific acoustic parameter adaptation data that is several orders of magnitude smaller than storage amounts otherwise required for a complete acoustic model. This provides customized acoustic speech models that increase recognition accuracy at a fraction of expected data storage requirements.Type: GrantFiled: May 22, 2014Date of Patent: March 3, 2015Assignee: Nuance Communications, Inc.Inventors: Vaibhava Goel, Peder A. Olsen, Steven J. Rennie, Jing Huang
-
Patent number: 8972256Abstract: A speech processing method and arrangement are described. A dynamic noise adaptation (DNA) model characterizes a speech input reflecting effects of background noise. A null noise DNA model characterizes the speech input based on reflecting a null noise mismatch condition. A DNA interaction model performs Bayesian model selection and re-weighting of the DNA model and the null noise DNA model to realize a modified DNA model characterizing the speech input for automatic speech recognition and compensating for noise to a varying degree depending on relative probabilities of the DNA model and the null noise DNA model.Type: GrantFiled: October 17, 2011Date of Patent: March 3, 2015Assignee: Nuance Communications, Inc.Inventors: Steven J. Rennie, Pierre Dognin, Petr Fousek
-
Publication number: 20140337026Abstract: A method and system for generating training data for a target domain using speech data of a source domain. The training data generation method including: reading out a Gaussian mixture model (GMM) of a target domain trained with a clean speech data set of the target domain; mapping, by referring to the GMM of the target domain, a set of source domain speech data received as an input to the set of target domain speech data on a basis of a channel characteristic of the target domain speech data; and adding a noise of the target domain to the mapped set of source domain speech data to output a set of pseudo target domain speech data.Type: ApplicationFiled: April 14, 2014Publication date: November 13, 2014Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATIONInventors: Osamu Ichikawa, Steven J. Rennie
-
Publication number: 20140257809Abstract: Techniques disclosed herein include using a Maximum A Posteriori (MAP) adaptation process that imposes sparseness constraints to generate acoustic parameter adaptation data for specific users based on a relatively small set of training data. The resulting acoustic parameter adaptation data identifies changes for a relatively small fraction of acoustic parameters from a baseline acoustic speech model instead of changes to all acoustic parameters. This results in user-specific acoustic parameter adaptation data that is several orders of magnitude smaller than storage amounts otherwise required for a complete acoustic model. This provides customized acoustic speech models that increase recognition accuracy at a fraction of expected data storage requirements.Type: ApplicationFiled: May 22, 2014Publication date: September 11, 2014Inventors: Vaibhava Goel, Peder A. Olsen, Steven J. Rennie, Jing Huang
-
Patent number: 8738376Abstract: Techniques disclosed herein include using a Maximum A Posteriori (MAP) adaptation process that imposes sparseness constraints to generate acoustic parameter adaptation data for specific users based on a relatively small set of training data. The resulting acoustic parameter adaptation data identifies changes for a relatively small fraction of acoustic parameters from a baseline acoustic speech model instead of changes to all acoustic parameters. This results in user-specific acoustic parameter adaptation data that is several orders of magnitude smaller than storage amounts otherwise required for a complete acoustic model. This provides customized acoustic speech models that increase recognition accuracy at a fraction of expected data storage requirements.Type: GrantFiled: October 28, 2011Date of Patent: May 27, 2014Assignee: Nuance Communications, Inc.Inventors: Vaibhava Goel, Peder A. Olsen, Steven J. Rennie, Jing Huang
-
Publication number: 20130096915Abstract: A speech processing method and arrangement are described. A dynamic noise adaptation (DNA) model characterizes a speech input reflecting effects of background noise. A null noise DNA model characterizes the speech input based on reflecting a null noise mismatch condition. A DNA interaction model performs Bayesian model selection and re-weighting of the DNA model and the null noise DNA model to realize a modified DNA model characterizing the speech input for automatic speech recognition and compensating for noise to a varying degree depending on relative probabilities of the DNA model and the null noise DNA model.Type: ApplicationFiled: October 17, 2011Publication date: April 18, 2013Applicant: NUANCE COMMUNICATIONS, INC.Inventors: Steven J. Rennie, Pierre Dognin, Petr Fousek