Li-Xuan Alex Peng

Electrical Engineering@National Tsing Hua University academic.alexpeng@gmail.com

profile_pic.jpg

Hinchu, Taiwan

Li-Xuan (Alex) Peng is a first-year master’s student at National Tsing Hua University (NTHU), Taiwan, advised by Prof. Min Sun and collaborate with Prof. Tsung-Wei Ke. His research focuses on physics-aware computer vision models and generative models for embodied vision.

Before joining NTHU, Alex was a research assistant at the Research Center for Information Technology Innovation (CITI), Academia Sinica, where he worked with Dr. Jun-Cheng Chen on adversarial attacks against generative diffusion models. He earned his B.S. in Electrical Engineering and Computer Science at National Central University (NCU). His academic excellence has been recognized with the Dr. Chu Shun-Yi Zyxel Scholarship, the most prestigious undergraduate award at NCU, as well as the Excellent Student Scholarship for outstanding academic performance.

Alex’s research interests span computer vision, generative models, explainable AI, CogSci-inspired AI, and neural dynamic models. He aspires to pursue a Ph.D. at the intersection of computer vision and cognitive neuroscience, aiming to uncover the fundamental principles of human visual intelligence.

Beyond research, Alex was a core member and instructor at the Google Developer Student Club (GDSC) at NCU, where he led workshops on machine learning, computer vision, and cloud-based AI training. His experience in mentoring students and organizing technical training reflects his commitment to fostering interdisciplinary research and knowledge-sharing in AI.

news

Dec 11, 2024 One paper has been accepted to AAAI 2025 :sparkles:
Nov 02, 2024 One paper has been accepted to WACV 2025 :sparkles:

selected publications

  1. AAAI 2025
    AtkPDM_manifold.png
    Pixel Is Not A Barrier: An Effective Evasion Attack for Pixel-Domain Diffusion Models
    Chun-Yen Shih*Li-Xuan Peng*, Jia-Wei Liao, Ernie Chu, Cheng-Fu Chou, and Jun-Cheng Chen
    Association for the Advancement of Artificial Intelligence (AAAI) 2025 , 2024
  2. WACV 2025
    two_stage_generation_pipeline.png
    DiffQRCoder: Diffusion-based Aesthetic QR Code Generation with Scanning Robustness Guided Iterative Refinement
    Jia-Wei Liao, Li-Xuan Peng*, Winston Wang*, Tzu-Sian Wang*, Ju-Hsuan Weng, Cheng-Fu Chou, and Jun-Cheng Chen
    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) , 2025