Taipei, Taipei City, Taiwan
An undergraduate researcher involved in AI for Healthcare & Trustworthy VLMs/LLMs.
BioMed-AI Lab, ECE/EECS, University of Michigan (Advisor: Prof. Liyue Shen) • Research on AI for healthcare, focusing on multi-turn clinical diagnostic reasoning with LLMs, knowledge graphs, and chain-of-reasoning-action consistency. [1]
FLAIR Lab, CSE, Texas A&M University (Advisor: Prof. Kuan-Hao Huang) • Research on vision-language alignment, interpretability, and hallucination in VLMs, focusing on cross-modal representations via sparse autoencoders.
AI Research Team, Far Eastern Memorial Hospital (Advisor: Dr. Fang-Ming Hung) • Research on explainable clinical decision support systems, integrating VLMs/LLMs, reinforcement learning, and knowledge graphs for disease/ICD prediction and medical plan generation. [1][2][3]
• Leveraged MLLM parallelization techniques to extract logistics data (e.g., MBL, invoices) via OCR and NLP, reducing processing time by ~67% (from 45s to 15s) while maintaining high accuracy in bounding box annotation and text-to-image mapping. • Built LLM-based web parsers with up to 97% accuracy, significantly reducing crawler maintenance for dynamic websites. • Presented research on LLM Agents and business use cases to support AI strategy.
• Core R&D contributor to LLM and Whisper projects, focused on transcription, inference acceleration, and LoRA-based model tuning. • Built RAG pipelines to enhance knowledge retrieval; developed LLM Agent for addiction counseling using structured dialogue planning.