Inventor · disambiguated record
Bita Darvish Rouhani
Also filed as: DARVISH ROUHANI BITA
6 granted patents·14 pending applications·5 citations·filing 2017–2024
70Inventor score
Top patents by PatentIndex Score
20 records- 0175US2025061320A1Adjusting activation compression for neural network trainingMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2024·Application pending·0 cites
- 0273US11676003B2Training neural network accelerators using mixed precision data formatsMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2018·Granted Jun 13, 2023·2 cites·23 claims
- 0372US12165038B2Adjusting activation compression for neural network trainingMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2019·Granted Dec 10, 2024·2 cites·21 claims
- 0465US11494614B2Subsampling training data during artificial neural network trainingMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2019·Granted Nov 8, 2022·1 cites·17 claims
- 0554US2023316042A1Mixture of experts models with sparsified weightsMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 0654US2024403644A1Optimizing deep neural network models based on sparsification and quantizationMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2023·Application pending·0 cites
- 0753US2023334284A1Sparsifying vectors for neural network models based on overlapping windowsMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 0852US12307372B2Data-aware model pruning for neural networksMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2021·Granted May 20, 2025·0 cites·17 claims
- 0952US12190235B2System for training an artificial neural networkMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2021·Granted Jan 7, 2025·0 cites·20 claims
- 1051US2023316080A1Sparsity masking methods for neural network trainingMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 1151US2023316039A1Sparsity and quantization for deep neural networksMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 1250US2023376725A1Model customization of transformers for improved efficiencyMICROSOFT TACHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 1349US2025355622A1Data scalingNVIDIA CORP·Filed 2024·Application pending·0 cites
- 1448US2022383092A1Turbo training for deep neural networksMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2021·Application pending·0 cites
- 1547US2024402993A1Determining shared exponent values for shared exponent floating point data typesMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2023·Application pending·0 cites
- 1645US2023385374A1Systems and methods for sparse matrix multiplicationMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2022·Application pending·0 cites
- 1742US11886833B2Hierarchical and shared exponent floating point data typesMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2021·Granted Jan 30, 2024·0 cites·20 claims
- 1841US2022405571A1Sparsifying narrow data formats for neural networksMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2021·Application pending·0 cites
- 1941US2020202213A1Scaled learning for training dnnMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2018·Application pending·0 cites
- 2039US2019197406A1Neural entropy enhanced machine learningMICROSOFT TECHNOLOGY LICENSING LLC·Filed 2017·Application pending·0 cites
Join the waitlist — get patent alerts
Get an alert when Bita Darvish Rouhani files or is granted a new patent.
We store only your email — no account needed. See our privacy policy.
Identity basis: PatentsView inventor disambiguation (2025Q4-odp release). How scoring works →