Patents

Patents (awarded)

  1. Context-sensitive search using a deep learning model,  U.S. Patent #9,535,960 granted on 1/3/2017
  2. Deep structured semantic model produced using click-through data, U.S. Patent #9,519,859 granted on 12/13/2016
  3. ·    Convolutional latent semantic models and their applications, U.S. Patent #9,477,654 granted on 10/25/2016
  4. ·    Computer-Implemented Deep Tensor Neural Network, U.S. Patent #9,292,787 granted on 3/22/2016
  5. ·    Discriminative pre-training of deep neural networks, U.S. Patent #9,235,799 granted on 1/12/2016
  6. ·    Tensor Deep Stacking Network, U.S. Patent #9,165,243, granted on October 20, 2015
  7. ·    Kernel deep convex networks and end-to-end learning, U.S. Patent #9,099,083, granted on August 4, 2015
  8. ·    Confidence calibration in automatic speech recognition systems, U.S. Patent #9,070,360, granted on June 30, 2015
  9. ·    Full-sequence training of deep structures for speech recognition, U.S. Patent #9,031,844, granted on May 12, 2015
  10. ·    Deep belief networks for large vocabulary continuous speech recognition, U.S. Patent #8,972,253, granted on March 3, 2015
  11. ·    Learning Processes For Single Hidden Layer Neural Networks With Linear Output Units, US Patent #8,918,352, granted on 12/23/2014
  12. ·    Exploiting Sparseness in Training Deep Neural Networks, filed 11/28/2011, US Patent #8700552, granted on 4/15/2014.
  13. ·    Online Distorted Speech Estimation Within An Unscented Transformation Framework. filed on 11/18/2010, US Patent #8731916, granted on 5/20/2014.
  14. ·    Deep Convex Network With Joint Use Of Nonlinear Random Projection, Restricted Boltzmann Machine And Batch-Based Parallelizable Optimization, filed 3/31/2011, US Patent #8489529, granted on 7/16/2013.
  15. ·    Deep structured conditional random fields for sequence labeling and classification, U.S. Patent; filed: 1/29/2010; granted on 6/25/2013, Patent #8,473,430.
  16. ·    Automatic reading feedback with parallel polarized language modeling,” (US Patent #8,433,576, granted on 4/30/2013
  17. ·    Generic framework for large-margin MCE training in speech recognition,” (US Patent #8,423,364, granted on 4/16/2013
  18. ·    Integrative and discriminative technique for spoken utterance translation (US Patent #8,407,041, granted on 3/26/2013
  19. ·    Speech recognition with non-liner noise reduction on Mel-frequency cepstra, (US Patent #8,306,817, granted Nov. 6, 2012)
  20. ·    Automatic Reading Tutoring, U.S. Patent; (US Patent #8,306,822, granted Nov. 6, 2012)
  21. ·    Adapting A Compressed Model For Use In Speech Recognition,” U.S. Patent, (#8,239,195, granted August 3, 2012)
  22. ·    Phase Sensitive Model Adaptation For Noisy Speech Recognition,” U.S. Patent, (#8,214,215, granted July 3, 2012)
  23. ·    Minimum classification error training with growth transformation optimization,” (U.S. Patent #8,301,449, granted Oct. 30, 2012)
  24. ·    Speech-centric multimodal user interface design in mobile technology,”  (US Patent #8,219,406, granted July 10, 2012)
  25. ·    High performance HMM adaptation with joint compensation of additive and convolutive distortions,” (US Patent #8,180,637, granted May. 15, 2012)
  26. ·    Piecewise-Based Variable-Parameter Hidden Markov Models and the Training Thereof,” (US Patent #8,160,878, granted April 17 2012)
  27. ·    Noise Suppressor for Robust Speech Recognition,” (US Patent #8,185,389, granted May. 22, 2012)
  28. ·    Parameter Clustering and Sharing for Variable-Parameter Hidden Markov Models, (US Patent #8,145,488, granted March 27, 2012)
  29. ·    Parameter Learning in Hidden Trajectory Model, (U.S. Patent #8,010,356, granted August 30, 2011)
  30. ·    Time Synchronous Decoding for Long-Span Hidden Trajectory Model, (US patent #7,877,256, granted 2011)
  31. ·    Integrated Speech Recognition and Semantic Classification (granted 2011, US patent #7,856,351)
  32. ·    Hidden Trajectory Modeling with Differential Cepstra for Speech Recognition, (granted 2010, US patent #7,805,308)
  33. ·    Segment-Discriminating Minimum Classification Error Pattern Recognition, with X. He and Q. Fu (granted Jan 18, 2011, US patent #7,873,209)
  34. ·    Hidden trajectory modeling with differential cepstra for speech recognition, U.S. Patent No.: 7,805,308; granted on September 28, 2010
  35. ·    Time Asynchronous Decoding for Long-Span Trajectory Model,” US patent No.: 7,734,460, granted on June 8, 2010
  36. ·    Method and Apparatus for Constructing a Speech Filter Using Estimates of Clean Speech and Noise,” U.S. Patent No.: 7,725,314; granted on May 25, 2010
  37. ·    Learning Statistically Characterized Resonance Targets in a Hidden Trajectory Model,  US patent #7653535, granted January 2010.
  38. ·    Incrementally Regulated Discriminative Margins in MCE Training for Speech Recognition, US patent #7617103, granted Sept 2009
  39. ·    Quantitative model for formant dynamics and contextually assimilated reduction in fluent speech, US patent No.: 7,565,292, granted on July 21, 2009
  40. ·    Acoustic models with structured hidden dynamics with integration over many possible hidden trajectories, US patent No.: 7,565,284, granted on July 21, 2009
  41. ·    Speaker-adaptive Learning of Resonance Targets in a Hidden Trajectory Model of Speech Coarticulation, US patent No.: 7,519,531, granted on April 14, 2009
  42. ·    Greedy algorithm for identifying values for vocal tract resonance vectors, U.S. Patent No.: 7,475,011; Granted on January 6, 2009
  43. ·    Method of Speech Recognition Using Multimodal Variational Inference with Switching State Space Models, U.S. Patent No.: 7,480,615; Granted on January 20, 2009
  44. ·    Method of Speech Recognition Using Variables Representing Dynamic Aspects of Speech, U.S. Patent No.: 7,346,510; Granted on March 18, 2008
  45. ·    Method of Noise Reduction Using Instantaneous Signal-to-Noise Ratio as the Principal Quantity for Optimal Estimation, U.S. Patent No.: 7,363,221; Granted on April 22, 2008
  46. ·    Method and Apparatus for Formant Tracking Using a Residual Model, U.S. Patent No.: 7,424,423; Granted on September 9, 2008
  47. ·    Multi-Sensory Speech Enhancement Using Synthesized Sensory Signal, U.S. Patent No.: 7,406,303; Granted on July 29, 2008
  48. ·    Two-stage implementation for phonetic recognition using a bi-directional target-directed model of speech co-articulation and reduction, U.S. Patent No.: 7,409,346; Granted on August 5, 2008
  49. ·    Removing noise from feature vectors, U.S. Patent No.: 7,310,599; Granted on December 18, 2007;
  50. ·    Method of determining uncertainty associated with acoustic distortion-based noise reduction, U.S. Patent No. 7,289,955; Granted on October 30, 2007
  51. ·    Method and apparatus for identifying noise environments from noisy signals, U.S. Patent No. 7,266,494; Granted on September 4, 2007
  52. ·    Method of noisy reduction using correction and scaling vectors with partitioning of the acoustic space in the domain of noisy speech, U.S. Patent No.7,254,536; Granted on August 7, 2007
  53. ·    Method of determining uncertainty in noise reduction, US and International Patents; U.S. Patent No.: 7,174,292; Granted on Feb. 6, 2007
  54. ·    Method of Noise Estimation Using Incremental Bayes Learning, US. Patent; Patent No.: 7,165,026; Granted on Jan. 16, 2007
  55. ·    Method of iterative noise estimation in a recursive framework, U.S. Patent; Patent No. 7,139,703; Granted on Nov. 21, 2006.
  56. ·    Method of noise reduction using correction vectors based on dynamic aspects of speech and noise normalization, United States Patent No. 7,117,148; Granted on October 3, 2006.
  57. ·    Method of noise reduction based on dynamic aspects of speech, United States Patent No. 7,107,210; Granted on Sept 12, 2006.
  58. ·    Method of pattern recognition using noise reduction uncertainty, United States Patent No. 7,103,540; Granted on Sept 5, 2006.
  59. ·    Microphone array signal enhancement using mixture models (jointly with Haggai Attias), United States Patent No. 7,103,541; Granted on Sept 5, 2006.
  60. ·    Efficient backward recursion for computing posterior probabilities, United States Patent No. 7,062,407; Granted on June 13, 2006.
  61. ·    Method of speech recognition using time-dependent interpolation and hidden dynamics, United States (and International) Patent No. 7,050,975; Granted on May 23, 2006.
  62. ·    Nonlinear observation models for removing noise from corrupted speech, United States (and International) Patent No. 7,047,047; Granted on May 16, 2006.
  63. ·    Method of Noise Reduction Using Correction and Scaling Vectors with Partitioning of the Acoustic Space in the Domain of Noisy Speech, United States Patent No. 7,003,455; Granted on February 21, 2006
  64. ·    Methods and Apparatus for Denoising and Dereverberation Using Variational Inference and Strong Speech Models, United States Patent No. 6,990,447; Granted on January 24, 2006
  65. ·    Method and Apparatus for Removing Noise from Feature Vectors, United States Patent No. 6,985,858; Granted on January 10, 2006
  66. ·    Methods for Including the Category of Environmental Noise When Processing Speech Signals, United States Patent No. 6,959,276; Granted on October 25, 2005
  67. ·    Method of iterative noise estimation in a recursive framework, United States Patent; Patent No. 6,944,590; Granted on September 13, 2005
  68. ·    Method of speech recognition using variational inference with switching state space models, United States Patent; Patent No. 6,931,374; Granted on August 16, 2005
  69. ·    Pattern Recognition Training Method and Apparatus Using Inserted Noise Followed by Noise Reduction, United States (and International) Patent; Patent No. 6,876,966; Granted on April 5, 2005
  70. ·    Apparatus for Speaker Clustering and for Speech Recognition, Patent No.: 2,965,537; Granted on Aug. 13, 1999; Countries of issue: United States and Japan.
  71. ·    Apparatus for Speaker Normalization Processor and for Voice Recognition Device, Patent No.: 2986792; Granted on Oct. 1, 1999; Countries of issue: United States and Japan.

 

Patents (Pending awards)

  1. ·    Method of speech recognition using hidden trajectory hidden Markov models, U.S. Patent
  2. ·    Zero-variance model of acoustic environment for enhancing noisy speech features,” U.S. Patent
  3. ·    Method and Apparatus for Multi-Sensory Speech Enhancement,” International Patent;
  4. ·    Method and apparatus for continuous valued vocal tract resonance tracking using piecewise linear approximation
  5. ·    Speech resonance target estimation using formant tracking results, U.S. Patent
  6. ·    Incrementally regulating discriminative margins in MCE training for speech recognition,” U.S. Patent; filing date: 8/25/2006
  7. ·    Using a discretized, higher order representation of hidden dynamic variables for speech recognition,” U.S. Patent; filing date: 8/21/2006
  8. ·    Integrated speech recognition and semantic classification,” U.S. Patent; filing date: 1/19/2007
  9. ·    Segment-discriminating minimum classification error pattern recognition,” U.S. Patent; filing date: 1/31/2007
  10. ·    Maximum Entropy Model with Continuous Features, U.S. Patent; filing date: April 2009
  11. ·    Cross-lingual speech recognition with HMM using KL distance,” U.S. Patent; filing date: April 2009
  12. ·    Maximum entropy model with continuous features, U.S. Patent; filing date: 4/1/2009
  13. ·    Discriminative learning of feature functions of generative type in speech translation, filed 10/28/2011
  14. ·    Discriminative pretraining of deep neural networks, filed 11/26/2011
  15. ·    Tensor Deep Stacking Networks, filed 2/15/2012.
  16. ·    Multilingual Deep Neural Network, filed 3/11/2013
  17. ·    Assignment of semantic labels to a sequence of words using neural network architectures, filed 9/2/2013
  18. ·    Deep structured semantic model produced using click-through data. filed 9/6/2013.
  19. ·    Convolutional Latent Semantic Models and Their Applications. filed 4/1/2014
  20. ·    Context-Sensitive Search Using a Deep Learning Model, filed 4/14/2014
  21. ·    Modeling Interestingness with Deep Neural Networks, filed 6/13/2014
  22. ·    Training and operations of computational models, US patent filed 6/29/2015
  23. ·    Leveraging global data for enterprise data analytics, US patent filed 7/24/2015
  24. ·    Representing learning using multi-task deep neural networks, US patent filed 7/28/2015
  25. ·    Semantically-relevant discovery of solutions, US patent filed 8/28/2015
  26. ·    Discovery of semantic similarities between images and text, US patent filed 8/28/2015
  27. Training and operating multi-layer computational models, US patent filed 11/23/2015
  28. ·    Multi-modal controller, US patent filed 12/30/2015
  29. ·    Multi-Stage Image Querying, filed with the U.S. Patent and Trademark Office on 4/12/2016.
  30. ·    Multiple action computational model training and operation, US patent filed 3/29/2016
  31. ·    Computational-model operation using multiple subject representations, US patent filed 3/29/2016
  32. ·    End-to-end memory networks for contextual language understanding, US patent filed 8/4/2016
  33. ·    Multi-domain joint semantic frame parsing, US patent filed 8/4/2016
  34. ·    Knowledge-guided structural attention processing, US patent filed 9/7/2016
  35. Joint language understanding and dialogue management, US patent filed 12/2/2016
  36. Systems and methods for automated query answer generation, US patent filed 12/2/2016
  37. End-to-End Learning of Dialogue Agents for Information Access, filed with U.S. Patent and Trademark Office on 1/13/2017
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s