Inventor · disambiguated record
George Andrei Saon
Also filed as: SAON GEORGE · SAON GEORGE A · SAON GEORGE ANDREI
31 granted patents·5 pending applications·357 citations·filing 1999–2023
96Inventor score
Top patents by PatentIndex Score
36 records- 0194US11783811B2Accuracy of streaming RNN transducerIBM·Filed 2020·Granted Oct 10, 2023·3 cites·19 claims
- 0294US6609093B1Methods and apparatus for performing heteroscedastic discriminant analysis in pattern recognition systemsIBM·Filed 2000·Granted Aug 19, 2003·101 cites·21 claims
- 0392US12136414B2Integrating dialog history into end-to-end spoken language understanding systemsIBM·Filed 2021·Granted Nov 5, 2024·3 cites·20 claims
- 0492US12046236B2Training end-to-end spoken language understanding systems with unordered entitiesIBM·Filed 2021·Granted Jul 23, 2024·3 cites·17 claims
- 0592US9665823B2Method and system for joint training of hybrid neural networks for acoustic modeling in automatic speech recognitionIBM·Filed 2014·Granted May 30, 2017·33 cites·16 claims
- 0691US11908454B2Integrating text inputs for training and adapting neural network transducer ASR modelsIBM·Filed 2021·Granted Feb 20, 2024·2 cites·20 claims
- 0789US10546575B2Using recurrent neural network for partitioning of audio data into segments that each correspond to a speech feature cluster identifierIBM·Filed 2016·Granted Jan 28, 2020·9 cites·20 claims
- 0888US10249292B2Using long short-term memory recurrent neural network for speaker diarization segmentationIBM·Filed 2016·Granted Apr 2, 2019·9 cites·20 claims
- 0988US7216077B1Lattice-based unsupervised maximum likelihood linear regression for speaker adaptationIBM·Filed 2000·Granted May 8, 2007·67 cites·4 claims
- 1084US9858919B2Speaker adaptation of neural network acoustic models using I-vectorsIBM·Filed 2014·Granted Jan 2, 2018·11 cites·10 claims
- 1181US11942078B2Chunking and overlap decoding strategy for streaming RNN transducers for speech recognitionIBM·Filed 2021·Granted Mar 26, 2024·1 cites·20 claims
- 1281US11908458B2Customization of recurrent neural network transducers for speech recognitionIBM·Filed 2020·Granted Feb 20, 2024·1 cites·19 claims
- 1377US12288551B2Accuracy of streaming RNN transducerIBM·Filed 2023·Granted Apr 29, 2025·0 cites·19 claims
- 1475US10468031B2Diarization driven by meta-information identified in discussion contentIBM·Filed 2017·Granted Nov 5, 2019·3 cites·14 claims
- 1574US11741946B2Multiplicative integration in neural network transducer models for end-to-end speech recognitionIBM·Filed 2020·Granted Aug 29, 2023·1 cites·16 claims
- 1672US9373324B2Applying speaker adaption techniques to correlated featuresIBM·Filed 2014·Granted Jun 21, 2016·4 cites·20 claims
- 1772US6385579B1Methods and apparatus for forming compound words for use in a continuous speech recognition systemIBM·Filed 1999·Granted May 7, 2002·65 cites·41 claims
- 1871US9704482B2Method and system for order-free spoken term detectionIBM·Filed 2015·Granted Jul 11, 2017·2 cites·20 claims
- 1970US7464031B2Speech recognition utilizing multitude of speech featuresIBM·Filed 2003·Granted Dec 9, 2008·17 cites·12 claims
- 2069US10902843B2Using recurrent neural network for partitioning of audio data into segments that each correspond to a speech feature cluster identifierIBM·Filed 2019·Granted Jan 26, 2021·1 cites·20 claims
- 2168US7529666B1Minimum bayes error feature selection in speech recognitionIBM·Filed 2000·Granted May 5, 2009·14 cites·10 claims
- 2267US11158303B2Soft-forgetting for connectionist temporal classification based automatic speech recognitionIBM·Filed 2019·Granted Oct 26, 2021·1 cites·19 claims
- 2365US10262260B2Method and system for joint training of hybrid neural networks for acoustic modeling in automatic speech recognitionIBM·Filed 2017·Granted Apr 16, 2019·1 cites·16 claims
- 2464US11120802B2Diarization driven by the ASR based segmentationIBM·Filed 2017·Granted Sep 14, 2021·1 cites·20 claims
- 2563US9378464B2Discriminative learning via hierarchical transformationsCASKEY SASHA P·Filed 2012·Granted Jun 28, 2016·3 cites·15 claims
- 2662US12387717B2Multi-speaker data augmentation for improved end-to-end automatic speech recognitionIBM·Filed 2023·Granted Aug 12, 2025·0 cites·20 claims
- 2762US9697830B2Method and system for order-free spoken term detectionIBM·Filed 2015·Granted Jul 4, 2017·1 cites·20 claims
- 2861US12482483B2Length perturbation techniques for improving generalization of deep neural network acoustic modelsIBM·Filed 2022·Granted Nov 25, 2025·0 cites·20 claims
- 2960US12444405B2Textual knowledge transfer for improved speech recognition and understandingIBM·Filed 2023·Granted Oct 14, 2025·0 cites·20 claims
- 3057US12148419B2Reducing exposure bias in machine learning training of sequence-to-sequence transducersIBM·Filed 2021·Granted Nov 19, 2024·0 cites·20 claims
- 3156US2025005348A1Neuronal activity modulation of artificial neural networksIBM·Filed 2023·Application pending·0 cites
- 3253US2024331687A1Insertion error reduction with confidence score-based word filteringIBM·Filed 2023·Application pending·0 cites
- 3351US2024169197A1Label smoothing technique for improving generalization of deep neural network acoustic modelsIBM·Filed 2022·Application pending·0 cites
- 3448US2008312921A1Speech recognition utilizing multitude of speech featuresAXELROD SCOTT E·Filed 2008·Application pending·0 cites
- 3545US2022319494A1End to end spoken language understanding modelIBM·Filed 2021·Application pending·0 cites
- 3640US11151996B2Vocal recognition using generally available speech-to-text systems and user-defined vocal trainingIBM·Filed 2019·Granted Oct 19, 2021·0 cites·20 claims
Identity basis: PatentsView inventor disambiguation (2025Q4-odp release). How scoring works →