Cheng-Che (SunnerLi) Lee 李政哲

Researcher

prof_pic_sunner_square.jpg
a6214123 (at) gmail (dot) com

I’m a researcher at Chunghwa Telecom in Taoyuan, Taiwan. I’m interested in making deep learning work where textbook assumptions break. Right now, I focus on small, efficient models for on-device AI agents and on keeping those agents secure. I also spent about three years working on generative models.

Before that, I spent five years at MediaTek, first as a computer vision and deep learning research engineer (data-driven RAW denoising for camera ISPs), then as an AI software engineer building LLM tools used in production. I received my M.S. from the Institute of Computer Science and Engineering at National Chiao Tung University, advised by Prof. Yu-Tai Ching and working with Prof. Wei-Chen Chiu and Dr. Li Su, and my B.S. from National Taipei University, where I worked with Prof. Bor-Shing Lin.

Outside of work, I ski ⛷️ and hike 🥾 abroad. See where I’ve been, and where I hope to go, in outside the lab. Feel free to reach out if you’d like to talk research, or if you’re looking for a ski or hiking buddy.

publications

  1. NeurIPS-W
    teaser_visual_routing.png
    Towards Lightweight Visual Routing for Mobile Agents via Fine-Grained Screenshot Synthesis
    Cheng-Che Lee, Chia-Pin Cheng, Hsu-Sheng Chang, and Chia-Lin Chou
    In NeurIPS 2026 Workshop on On-Device Intelligence: Foundation Models under Real-World Constraints , 2026
    PDF Website
  2. NeurIPS-W
    teaser_sms_homoglyph.png
    Beyond Tool Accuracy: End-to-End Reliability of On-Device SMS Agents Under Homoglyph Attacks
    Cheng-Che Lee, Chia-Pin Cheng, Hsu-Sheng Chang, and Chia-Lin Chou
    In NeurIPS 2026 Workshop on Agents in the Wild: Safety, Security, and Beyond , 2026
    PDF Website
  3. ICPR
    teaser_low.png
    Learning Low-Shot Generative Networks for Cross-Domain Data
    Hsuan-Kai Kao, Cheng-Che Lee, and Wei-Chen Chiu
    In 25th International Conference on Pattern Recognition (ICPR) , 2020
    PDF Code
  4. ACM MM
    teaser_cross.png
    Crossing You in Style: Cross-modal Style Transfer from Music to Visual Arts
    Cheng-Che Lee*, Wan-Yi Lin*, Yen-Ting Shih, Pei-Yi Patricia Kuo, and Li Su
    In Proceedings of the 28th ACM international conference on multimedia (MM) , 2020
    PDF Code Website

other contributions

  1. NeurIPS-W
    teaser_soc.png
    Fast SoC thermal simulation with physics-aware U-Net
    Yu-Sheng Lin, Li-Song Lin, Chin-jui Chang, Ting-Yu Lin, Ya-Wen Yu, and 6 more authors
    In Neural Information Processing Systems (NeurIPS) workshop on Machine Learning and the Physical Sciences , 2023
    Acknowledged contributor, not an author: I helped run experiments and supported the paper writing. Separately, on the same MediaTek project, I proposed a ranking-based training objective that improved floorplan ranking accuracy by over 50%.
    PDF

outside the lab

Skiing, hiking and wandering, starting from Taiwan.