Dr. Yanchen Li

Postdoctoral Researcher

Advancing research in deep learning with a focus on compression and acceleration techniques that optimize neural network performance and efficiency.

View Publications

About Me

Research Background & Experience

Research Focus

I'm a postdoctoral researcher specializing in deep learning. My research focuses on addressing challenging problems in accelerating DNN models with low-precision quantization and pruning techniques.

Currently, I'm working at RIKEN Center for Computer Science, where I am a member of the Advanced AI Device Development Unit leaded by Prof. Kentaro Sano to develop novel AI accelerators that push the boundaries of computational efficiency.

I received my Ph.D. in Computer Science from the University of Osaka in 2025, where I worked under the supervision of Prof. Fumihiko Ino on accelerating DNN models on GPUs and developing innovative optimization strategies.

Contact Me

Publications

Recent Research Contributions

A Theoretical and Empirical Exploration of TileTrans for Effective Tile Pruning

Yanchen Li, Fumihiko Ino

Knowledge-Based Systems, 2024/8/8

Enhancing Bayesian Optimization with NLML-Based Transfer Learning

Yanchen Li, Taku Tsuzuki

2024 IEEE 20th International Conference on e-Science (e-Science), 2024/9/16

Parallel Heuristic Methods to Accelerate Best Equivocation Code Generation

Yanchen Li, Ke Zhang, Fumihiko Ino

IEEE Access, 2023/5/3

A One-Shot Reparameterization Method for Reducing the Loss of Tile Pruning on DNNs

Yanchen Li, Qingzhong Ai, Fumihiko Ino

International Joint Conference on Neural Networks (IJCNN), 2022/7/18

Awards & Honors

Recognition & Achievements

Excellent Student Presentation Award

Summer United Workshops on Parallel, Distributed and Cooperative Processing, 2024 (SWoPP 2024)

Humanware Innovation Program

Osaka University, 2019-2025

Information AI Fellowship

Osaka University, 2022-2025

Contact

Get In Touch

Email

yanchen.li@riken.jp

Office Location

〒650-0047 Hyogo, Kobe, Chuo Ward, Minatojima Minamimachi, 6 Chome-3-5, 210