研究キーワード (1件):
AI Processor,Deep Neural Network, STT-MRAM, Pedestrian Detection
競争的資金等の研究課題 (3件):
2021 - 2024 A Novel Power Reduction Technique Using Error-resilient Deep Neural Networks for STT-MRAM Based Energy-efficient Brain-inspired Processor Design
Li Zhang, Tao Li, Tetsuo Endoh. Small Area and High Throughput Error Correction Module of STT-MRAM for Object Recognition Systems. IEEE Transactions on Industrial Informatics. 2024. 20. 5. 7777-7786
Tao Li, Li Zhang, Yitao Ma, Tetsuo Endoh. Bridging Artificial Intelligence and Devices: Power Reduction Method of Non-volatile Devices with Error-resilient Deep Neural Networks. IEEE Transactions on Magnetics. 2023. 1-9
Tao Li, Yitao Ma, Ko Yoshikawa, Tetsuo Endoh. Erratum: Hybrid Signed Convolution Module With Unsigned Divide-and-Conquer Multiplier for Energy-Efficient STT-MRAM-Based AI Accelerator (IEEE Transactions on Very Large Scale Integration (VLSI) Systems (2023) DOI: 10.1109/TVLSI.2023.3245099). IEEE Transactions on Very Large Scale Integration (VLSI) Systems. 2023. 31. 6. 906
Tao Li, Yitao Ma, Tetsuo Endoh. Neuromorphic processor-oriented hybrid Q-format multiplication with adaptive quantization for tiny YOLO3. Neural Computing and Applications. 2023. 35. 15. 11013-11041
Tao Li, Yitao Ma, Ko Yoshikawa, Tetsuo Endoh. Hybrid Signed Convolution Module with Unsigned Divide-and-Conquer Multiplier for Energy-efficient STT-MRAM-based AI Accelerator. IEEE Transactions on Very Large Scale Integration (VLSI) Systems. 2023. 31. 7. 1-5