Patents/Inventors

Eric S. Chung

Woodinville, WAUS33 patents

26 Patents

US126082252026
Hierarchical Programming Model for Artificial Intelligence Hardware
Microsoft Technology Licensing, LLC
0 cites
US125859262026
Adjusting Precision and Topology Parameters for Neural Network Training Based on a Performance Metric
Microsoft Technology Licensing, LLC
0 cites
US125669622026
Dithered Quantization of Parameters During Training with a Machine Learning Tool
Microsoft Technology Licensing, LLC
0 cites
US125544662026
Systems and Methods for Accelerating the Computation of the Exponential Function
Microsoft Technology Licensing, LLC
0 cites
US125478722026
Machine Learning Model Processing Based on Perplexity
Microsoft Technology Licensing, LLC
0 cites
US125113442025
Fusing Operators for Neural Network Hardware Accelerators
Microsoft Technology Licensing, LLC
0 cites
US124438482025
Neural Network Activation Compression with Narrow Block Floating-point
Microsoft Technology Licensing, LLC
0 cites
US124120882025
Reducing Operations for Training Neural Networks
Microsoft Technology Licensing, LLC
0 cites
US123462772025
Systems and Methods for Hardware Acceleration of Data Masking
Microsoft Technology Licensing, LLC
0 cites
US123073552025
Neural Network Processing with Chained Instructions
Microsoft Technology Licensing, LLC
0 cites
US123073722025
Data-aware Model Pruning for Neural Networks
Microsoft Technology Licensing, LLC
0 cites
US122775022025
Neural Network Activation Compression with Non-uniform Mantissas
Microsoft Technology Licensing, LLC
0 cites
US121650382024
Adjusting Activation Compression for Neural Network Training
Microsoft Technology Licensing, LLC
0 cites
US120674952024
Neural Network Activation Compression with Non-uniform Mantissas
Microsoft Technology Licensing, LLC
0 cites
US120457242024
Neural Network Activation Compression with Outlier Block Floating-point
Microsoft Technology Licensing, LLC
0 cites
US119343272024
Systems and Methods for Hardware Acceleration of Data Masking Using a Field Programmable Gate Array
Microsoft Technology Licensing, LLC
0 cites
US118868332024
Hierarchical and Shared Exponent Floating Point Data Types
Microsoft Technology Licensing, LLC
0 cites
US117902122023
Quantization-aware Neural Architecture Search
MICROSOFT TECHNOLOGY LICENSING, LLC
0 cites
US116760032023
Training Neural Network Accelerators Using Mixed Precision Data Formats
Microsoft Technology Licensing, LLC
0 cites
US116634502023
Neural Network Processing with Chained Instructions
Microsoft Technology Licensing, LLC
0 cites
US116454932023
Flow for Quantized Neural Networks
Microsoft Technology Licensing, LLC
0 cites
US116049602023
Differential Bit Width Neural Architecture Search
Microsoft Technology Licensing, LLC
0 cites
US115868832023
Residual Quantization for Neural Networks
Microsoft Technology Licensing, LLC
0 cites
US115742392023
Outlier Quantization for Training and Inference
MICROSOFT TECHNOLOGY LICENSING, LLC
0 cites
US115622472023
Neural Network Activation Compression with Non-uniform Mantissas
Microsoft Technology Licensing, LLC
0 cites
US115567622023
Neural Network Processor Based on Application Specific Synthesis Specialization Parameters
Microsoft Technology Licensing, LLC
0 cites