|
PAT. NO.
|
|
Title
|
| 10,354,656 (opens in new tab) |
 |
Speaker recognition (opens in new tab) |
| 10,347,241 (opens in new tab) |
|
Speaker-invariant training via adversarial learning (opens in new tab) |
| 10,235,994 (opens in new tab) |
|
Modular deep learning model (opens in new tab) |
| 10,115,393 (opens in new tab) |
|
Reduced size computerized speech model speaker adaptation (opens in new tab) |
| 10,019,990 (opens in new tab) |
|
Variable-component deep neural network for robust speech recognition (opens in new tab) |
| 10,019,984 (opens in new tab) |
|
Speech recognition error diagnosis (opens in new tab) |
| 9,997,161 (opens in new tab) |
|
Automatic speech recognition confidence classifier (opens in new tab) |
| 9,842,585 (opens in new tab) |
|
Multilingual deep neural network (opens in new tab) |
| 9,728,184 (opens in new tab) |
|
Restructuring deep neural network acoustic models (opens in new tab) |
| 9,520,127 (opens in new tab) |
|
Shared hidden layer combination for speech recognition systems (opens in new tab) |
| 9,454,958 (opens in new tab) |
|
Exploiting heterogeneous data in deep neural network-based speech recognition systems (opens in new tab)
|
| 9,336,775 (opens in new tab) |
|
Posterior-based feature with partial distance elimination for speech recognition (opens in new tab) |
| 9,324,321 (opens in new tab) |
|
Low-footprint adaptation and personalization for a deep neural network (opens in new tab) |
| 9,280,969 (opens in new tab) |
|
Model training for automatic speech recognition from imperfect transcription data (opens in new tab) |
| 9,263,030 (opens in new tab) |
|
Adaptive online feature normalization for speech recognition (opens in new tab) |
| 9,208,777 (opens in new tab) |
|
Feature space transformation for personalization using generalized i-vector clustering (opens in new tab) |
| 8,731,916 (opens in new tab) |
|
Online distorted speech estimation within an unscented transformation framework (opens in new tab) |
| 8,700,400 (opens in new tab) |
|
Subspace speech adaptation (opens in new tab) |
| 8,457,946 (opens in new tab) |
|
Recognition architecture for generating Asian characters (opens in new tab) |
| 8,306,819 (opens in new tab) |
|
Enhanced automatic speech recognition using mapping between unsupervised and supervised speech model parameters trained on same acoustic training data (opens in new tab) |
| 8,239,195 (opens in new tab) |
|
Adapting a compressed model for use in speech recognition (opens in new tab) |
| 8,214,215 (opens in new tab) |
|
Phase sensitive model adaptation for noisy speech recognition (opens in new tab) |
| 8,185,389 (opens in new tab) |
|
Noise suppressor for robust speech recognition (opens in new tab) |
| 8,185,376 (opens in new tab) |
|
Identifying language origin of words (opens in new tab) |
| 8,180,637 (opens in new tab) |
|
High performance HMM adaptation with joint compensation of additive and convolutive distortions (opens in new tab) |
| 8,160,878 (opens in new tab) |
|
Piecewise-based variable-parameter Hidden Markov Models and the training thereof (opens in new tab) |
| 8,145,488 (opens in new tab) |
|
Parameter clustering and sharing for variable-parameter hidden markov models (opens in new tab) |
| 8,086,455 (opens in new tab) |
|
Model development authoring, generation and execution based on data and processor dependencies (opens in new tab) |
| 7,860,716 (opens in new tab) |
|
Speech model refinement with transcription error detection (opens in new tab) |
| 7,516,069 (opens in new tab) |
|
Middle-end solution to robust speech recognition (opens in new tab) |
| 7,464,033 (opens in new tab) |
|
Decoding multiple HMM sets using a single sentence grammar (opens in new tab) |
| 7,451,082 (opens in new tab) |
|
Noise-resistant utterance detector (opens in new tab) |
| 7,269,558 (opens in new tab) |
|
Decoding multiple HMM sets using a single sentence grammar (opens in new tab) |
| 7,236,930 (opens in new tab) |
|
Method to extend operating range of joint additive and convolutive compensating algorithms (opens in new tab) |
| 7,165,028 (opens in new tab) |
|
Method of speech recognition resistant to convolutive distortion and additive distortion (opens in new tab) |
| 7,103,547 (opens in new tab) |
|
Implementing a high accuracy continuous speech recognizer on a fixed-point processor (opens in new tab) |
| 7,089,183 (opens in new tab) |
|
Accumulating transformations for hierarchical linear regression HMM adaptation (opens in new tab) |
| 7,062,433 (opens in new tab) |
|
Method of speech recognition with compensation for both channel distortion and background noise (opens in new tab) |
| 6,980,952 (opens in new tab) |
|
Source normalization training for HMM modeling of speech (opens in new tab) |
| 6,980,950 (opens in new tab) |
|
Automatic utterance detector with high noise immunity (opens in new tab) |
| 6,912,497 (opens in new tab) |
|
Calibration of speech data acquisition path (opens in new tab) |
