Harshit Khaitan
32Patents
8h-index
15Co-inventors
64Inventor score
Filing activity: Oct 27, 2016 → Feb 21, 2023
Most-cited inventions
| Patent | Title | Area | Cited by | Status |
|---|---|---|---|---|
| US9710265B1 | Neural network compute tile | Physics | 110 | Active |
| US9836691B1 | Neural network instruction set architecture | Physics | 76 | Active |
| US10175980B2 | Neural network compute tile | Physics | 34 | Active |
| US10504022B2 | Neural network accelerator with parameters resident on chip | Physics | 19 | Active |
| US9959498B1 | Neural network instruction set architecture | Physics | 19 | Active |
| US10534607B2 | Accessing data in multi-dimensional tensors using adders | Physics | 10 | Active |
| US10496326B2 | Hardware double buffering using a special purpose computational unit | Physics | 9 | Active |
| US10802956B2 | Accessing prologue and epilogue data | Physics | 8 | Active |
| US11709783B1 | Tensor data distribution using grid direct-memory access (DMA) controller | Physics | 8 | Active |
| US9946539B1 | Accessing data in multi-dimensional tensors using adders | Physics | 8 | Active |
| US10248908B2 | Alternative loop limits for accessing data in multi-dimensional tensors | Physics | 7 | Active |
| US11467675B1 | Multi-component detection of gestures | Physics | 4 | Active |
| US10108538B1 | Accessing prologue and epilogue data | Physics | 3 | Active |
| US10175912B1 | Hardware double buffering using a special purpose computational unit | Physics | 2 | Active |
| US11704562B1 | Architecture for virtual instructions | Physics | 2 | Active |
| US11422801B2 | Neural network compute tile | Physics | 1 | Active |
| US11176493B2 | Virtualizing external memory as local to a machine learning accelerator | Emerging Cross-Sectional Technologies | 1 | Active |
| US12001893B1 | Distributed synchronization scheme | Physics | 1 | Active |
| US10885434B2 | Alternative loop limits for accessing data in multi-dimensional tensors | Physics | 0 | Active |
| US11893159B2 | Multi-component detection of gestures | Physics | 0 | Active |
| US11501144B2 | Neural network accelerator with parameters resident on chip | Physics | 0 | Active |
| US12061968B2 | Neural network instruction set architecture | Physics | 0 | Active |
| US11099772B2 | Hardware double buffering using a special purpose computational unit | Physics | 0 | Active |
| US11954580B2 | Spatial tiling of compute arrays with shared control | Physics | 0 | Active |
| US11922306B2 | Tensor controller architecture | Physics | 0 | Active |
Source: USPTO / EPO open patent data. Inventor disambiguation is heuristic; counts are objective bibliographic measures.