About

I am a Ph.D. student at National Taiwan University (NTU), advised by Prof. Hung-yi Lee and Prof. Jyh-Shing Roger Jang. I received my M.S. degree from NTU in 2023 and B.S. degree from Taiwan Tech in 2020. My research interests include, Speech Processing, Audio/Text Large Language Models (LLMs), and Audio Deepfakes.

I have focused on audio deepfake detection from both singing and speech perspectives. I proposed SingGraph to bridge music and speech understanding for SingFake detection, and developed CodecFake+, a large-scale dataset using neural codec resynthesis as a scalable proxy for TTS/VC data. Furthermore, we introduced the source tracing task along with SASTNet, recognized as a 🏆 Best Student Paper nominee at IEEE ASRU 2025.

Recently, I have expanded my research scope to Audio/Text LLMs, including contributions to DeSTA2.5, Codec-SUPERB, and Dynamic-SUPERB Phase 2. Meanwhile, I also mentor research on RAG; our work "RAG for Historical Archives," received the 🏆 Best Paper Award at ROCLING 2025.

Selected Publications (* equal contribution)

Experience

Selected Honors

Selected Services