Yi LI

Yi LI

Postdoctoral Fellow

liyi126@hku.hk
+85253091843
Room324, HW
Personal Page Google Scholar

Biography

Dr. Yi Li is currently a Post-doctoral Fellow in the Department of Electrical and Computer Engineering (ECE) at the University of Hong Kong (HKU). His research focuses on the algorithm-hardware co-design for artificial intelligence based on emerging memories, such as ReRAM and the development of Compute-in-Memory (CIM) accelerators. He received his Ph.D. degree in Microelectronics and Solid-State Electronics from the Institute of Microelectronics, Chinese Academy of Sciences (IMECAS). Prior to that, he earned his Bachelor’s and Master’s degrees from Xi’an Jiaotong University. During his Ph.D. and post-doctoral research, Dr. Li has conducted in-depth studies on the application of emerging memories in intelligent computing. His primary research directions include: 1) developing ReRAM-based multimodal event data learning circuits to enable continuous learning, thereby promoting the intelligence and low-power operation of edge AI applications; 2) proposing mixed-precision compute-in-memory neural network architectures to balance system energy efficiency and inference accuracy; and 3) building a full-stack co-optimization framework spanning from devices and circuits to systems and algorithms, achieving cross-layer performance enhancement and verification. Dr. Li’s research findings have been widely published in top-tier international conferences and journals, including DAC, IEDM, Nature Machine Intelligence, Nature Computational Science, Nature Communications, and Advanced Materials. Currently, he is dedicated to advancing the practical applications of emerging memory and compute-in-memory technologies in edge intelligence, hardware security, and novel neuromorphic computing systems.

Publications

2026

  • A Fully Analog Continuous-Time CIM Neural ODE Solver for Flow-Matching-Based Fluid Dynamics GenerationSongqi Wang, Meng Xu, Jichang Yang, Zhexu Chen, Hegan Chen, Sishuo Liu, Xinyuan Zhang, Kwun Hang WONG, Ning Lin, Yi Li, Zhongrui Wang and Han Wang2026 63rd ACM/IEEE Design Automation Conference (DAC)C.
  • Reconfigurable 3D-VRRAM LUT-CiM with Greedy G-shuffle Quantization for End-to-End 4-bit ViTs in Edge Dense PredictionYi Li, Zijian Ye, Zhaori Cong, Shengzhe Yan, Songqi Wang, Ning Lin, Jinshan Yue, xiaojuan qi, Zhongrui Wang and Han Wang2026 63rd ACM/IEEE Design Automation Conference (DAC)C.
  • Pruning random resistive memory for optimizing analog AILi, Yi and Wang, Songqi and Zhao, Yaping and Wang, Shaocong and Wang, Bo and Zhang, Woyu and He, Yangu and Lin, Ning and Cui, Binbin and Chen, Xi and othersNature CommunicationsJ.
  • A Physics-Prior Intelligent Compact Modeling Framework for BEOL-Compatible DTCO: EKAN-based Distillation from Neural to Symbolic ModelsXufan Li, Yi Li, Ning Lin, Xiaoyi Zhang, Yue Zhao, Jiaye Shen, xiaojuan qi, Zhongrui Wang, Han Wang, Zhenjie Yao, Lingfei Wang, ling li and Ming Liu2026 63rd ACM/IEEE Design Automation Conference (DAC)C.

2025

  • Continuous-time digital twin with analog memristive neural ordinary differential equation solverChen, Hegan and Yang, Jichang and Chen, Jia and Wang, Songqi and Wang, Shaocong and Wang, Dingchen and Tian, Xinyu and Yu, Yifei and Chen, Xi and Lin, Yinan and othersScience AdvancesJ.
  • Topology optimization of random memristors for input-aware dynamic SNNWang, Bo and Zhang, Xinyuan and Wang, Shaocong and Lin, Ning and Li, Yi and Yu, Yifei and Zhang, Yue and Yang, Jichang and Wu, Xiaoshan and He, Yangu and othersScience AdvancesJ.
  • Random resistive memory-based deep extreme point learning machine for unified visual processingWang, Shaocong and Gao, Yizhao and Li, Yi and Zhang, Woyu and Yu, Yifei and Wang, Bo and Lin, Ning and Chen, Hegan and Zhang, Yue and Jiang, Yang and othersNature CommunicationsJ.
  • Resistive memory-based zero-shot liquid state machine for multimodal event data learningLin, Ning and Wang, Shaocong and Li, Yi and Wang, Bo and Shi, Shuhui and He, Yangu and Zhang, Woyu and Yu, Yifei and Zhang, Yue and Zhang, Xinyuan and othersNature Computational ScienceJ.
  • Efficient modeling of ionic and electronic interactions by a resistive memory-based reservoir graph neural networkXu, Meng and Wang, Shaocong and He, Yangu and Li, Yi and Zhang, Woyu and Yang, Ming and Qi, Xiaojuan and Wang, Zhongrui and Xu, Ming and Shang, Dashan and othersNature Computational ScienceJ.
  • Efficient Edge Vision Transformer Accelerator with Decoupled Chunk Attention and Hybrid Computing-In-MemoryLi Yi and Ye Zijian and Fu Xiangqu and Wang Songqi and Du Shucheng and Lin Ning and Shang Dashan and Yue Jinshan and Wang Zhongrui and Qi Xiaojuan and Zhang Feng and Wang Han2025 62nd ACM/IEEE Design Automation Conference (DAC)C.
  • Brain-inspired in-memory data pruning and computing with TaOx Mem-SelectorsLi, Yi and Lai, Jinru and Wang, Songqi and Lin, Ning and Zheng, Xu and Sun, Wenxuan and Dong, Danian and Xu, Xiqing and Ma, Haili and Zhang, Feng and othersAdvanced MaterialsJ.
  • Re 4 PUF: A Reliable, Reconfigurable ReRAM-based PUF Resilient to DNN and Side Channel AttacksLin, Ning and Li, Yi and He, Yangu and Wang, Songqi and Chen, Hegan and Wong, Kwunhang and Li, Chuxin and Yang, Jichang and Yu, Yifei and Xu, Meng and others2025 62nd ACM/IEEE Design Automation Conference (DAC)C.
  • Guarder: A Stable and Lightweight Reconfigurable RRAM-based PIM Accelerator for DNN IP ProtectionLin, Ning and Li, Yi and Li, Jiankun and Yang, Jichang and He, Yangu and Luo, Yukui and Shang, Dashan and Chen, Xiaoming and Qi, Xiaojuan and Wang, Zhongrui2025 62nd ACM/IEEE Design Automation Conference (DAC)C.

2024

  • CMN: a co-designed neural architecture search for efficient computing-in-memory-based mixture-of-expertsHan, Shihao and Liu, Sishuo and Du, Shucheng and Li, Mingzi and Ye, Zijian and Xu, Xiaoxin and Li, Yi and Wang, Zhongrui and Shang, DashanScience China Information SciencesJ.
  • Fully Binarized Graph Convolutional Network Accelerator Based on In-Memory Computing with Resistive Random-Access MemoryZhang, Woyu and Li, Zhi and Zhang, Xinyuan and Wang, Fei and Wang, Shaocong and Lin, Ning and Li, Yi and Wang, Jun and Yue, Jinshan and Dou, Chunmeng and othersAdvanced Intelligent SystemsJ.

2023

  • A scalable small-footprint time-space-pipelined architecture for reservoir computingDai, Zhuoyu and Xiang, Feibin and He, Chaojie and Wang, Zi and Zhang, Woyu and Li, Yi and Yue, Jinshan and Shang, DashanIEEE Transactions on Circuits and Systems II: Express BriefsJ.
  • Echo state graph neural networks with analogue random resistive memory arraysWang, Shaocong and Li, Yi and Wang, Dingchen and Zhang, Woyu and Chen, Xi and Dong, Danian and Wang, Songqi and Zhang, Xumeng and Lin, Peng and Gallicchio, Claudio and othersNature Machine IntelligenceJ.
  • An ADC-less RRAM-based computing-in-memory macro with binary CNN for efficient edge AILi, Yi and Chen, Jia and Wang, Linfang and Zhang, Woyu and Guo, Zeyu and Wang, Jun and Han, Yongkang and Li, Zhi and Wang, Fei and Dou, Chunmeng and othersIEEE Transactions on Circuits and Systems II: Express BriefsJ.
  • Point-of-care testing (POCT) system based on self-recovery memoristor chip with low energy consuption (1.547 TOPS/W) and high recognition (1142 fram/s)Zheng, Xu and Wu, Lizhou and Liu, Yixuan and Wu, Qiqiao and Xie, Yuanlu and Li, Yi and Lai, Jinru and Sun, Wenxuan and Dong, Danian and Yu, Jie and others2023 International Electron Devices Meeting (IEDM)C.
  • Bioinspired in-sensor reservoir computing for self-adaptive visual recognition with two-dimensional dual-mode phototransistorsJiang, Nanjia and Tang, Jian and Zhang, Woyu and Li, Yi and Li, Na and Li, Xiuzhen and Chen, Xi and Fang, Renrui and Guo, Zeyu and Wang, Fei and othersAdvanced Optical MaterialsJ.
  • High area efficiency (6 TOPS/mm 2) multimodal neuromorphic computing system implemented by 3D multifunctional RRAM arraySun, Wenxuan and Li, Yi and Zhang, Woyu and Zheng, Xu and Dong, Danian and Yu, Jie and Lai, Jinru and Fan, Shaoyang and Wang, Hongzhou and Xu, Xiaoxin and others2023 International Electron Devices Meeting (IEDM)C.

2022

  • Few-shot graph learning with robust and energy-efficient memory-augmented graph neural network (MAGNN) based on homogeneous computing-in-memoryZhang, Woyu and Wang, Shaocong and Li, Yi and Xu, Xiaoxin and Dong, Danian and Jiang, Nanjia and Wang, Fei and Guo, Zeyu and Fang, Renrui and Dou, Chunmeng and others2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits)C.
  • 3d reservoir computing with high area efficiency (5.12 tops/mm 2) implemented by 3d dynamic memristor array for temporal signal processingSun, Wenxuan and Zhang, Woyu and Yu, Jie and Li, Yi and Guo, Zeyu and Lai, Jinru and Dong, Danian and Zheng, Xu and Wang, Fei and Fan, Shaoyang and others2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits)C.
  • Convolutional Echo-State Network with Random Memristors for Spatiotemporal Signal ClassificationWang, Shaocong and Chen, Hegan and Zhang, Woyu and Li, Yi and Wang, Dingchen and Shi, Shuhui and Zhao, Yaping and Loong, Kam Chi and Chen, Xi and Dong, Yujiao and othersAdvanced Intelligent SystemsJ.
  • Mixed-Precision Continual Learning Based on Computational Resistance Random Access MemoryLi, Yi and Zhang, Woyu and Xu, Xiaoxin and He, Yifan and Dong, Danian and Jiang, Nanjia and Wang, Fei and Guo, Zeyu and Wang, Shaocong and Dou, Chunmeng and othersAdvanced Intelligent SystemsJ.

2021

  • Energy efficient and robust reservoir computing system using ultrathin (3.5 nm) ferroelectric tunneling junctions for temporal data learningYu, Jie and Li, Yi and Sun, Wenxuan and Zhang, Woyu and Gao, Zhaomeng and Dong, Danian and Yu, Zhaoan and Zhao, Yulin and Lai, Jinru and Ding, Qingting and others2021 Symposium on VLSI TechnologyC.