陳飛岳 Fei-Yueh Chen

Education學歷

• University of Rochester Aug 2024 – May 2026

  M.S. in Electrical Engineering & M.S. in Computational Linguistics (Dual Degree)

  Rochester, New York

• National Taiwan University Sep 2017 – Jun 2022

  B.A. in Economics, Minor in Mathematics

  Taipei, Taiwan  ·  Dean's List Award (2019)

Industry業界經歷

• Smule Inc. May 2025 – Dec 2025

  Research Intern  ·  Remote, USA

  Advisor: Yongyi Zang

  • Contributed to the dataset and evaluation pipeline for Smule Renaissance Small (SRS), a 10.4M-parameter vocal restoration model achieving 10.5× real-time inference on iPhone 12 CPU at 48 kHz
  • Contributed to StylePitcher: rectified flow matching-based pitch curve generator for singing voice synthesis — NeurIPS 2025 Workshop Best Paper Award
  • Trained HuBERT-based self-supervised encoders and distilled ConvNeXt-based speech encoders for on-device singing voice applications

Research研究經歷

• AIR Lab, University of Rochester Aug 2024 – Present

  Graduate Research Student

  Advisor: Prof. Zhiyao Duan

  • Co-developed MusicSem: 32K-pair language-audio dataset from Reddit with 5-category semantic taxonomy (NeurIPS 2025 Workshop)
  • Developing a hierarchical method for symbolic music generation combining autoregressive and diffusion models
  • Collaborated with Prof. C. M. Downey on evaluating multilingual ASR for Yoruba and Uneme; extended Feature Error Rate with tone-aware dimensions (TER) (EACL 2026 AfricaNLP)
• CCL Lab, National Taiwan University Jun 2024 – Aug 2024

  Research Intern

  Advisors: Prof. Kuan-Ming Chen, Prof. Patrick C.-C. Huang

  • Built LLM pipeline for structured feature extraction from Mandarin court verdicts, enabling large-scale empirical legal analysis
  • Co-developing TW-LegalBench: 16K+ MCQs, 117 essay questions, 14K+ criminal judgment prediction instances (under review, ICAIL 2026)
• NLPSA Lab, Academia Sinica Sep 2022 – Jun 2024

  Research Assistant

  Advisor: Prof. Lun-Wei Ku

  • Led development of Omiyage: LLM-powered multimodal interactive art system — Taiwan Future Tech Award 2023. Deployed across 4 public exhibitions
• Music AI Lab, Academia Sinica Sep 2021 – Aug 2022

  Research Assistant

  Advisor: Prof. Yi-Hsuan Yang

  • Proposed novel MIDI tokenization for transformer-based symbolic music generation enabling joint lyrics-melody generation in a single autoregressive pass
  • Improved lyric rhyming fluency via 2-stage generation with phoneme-level tone and vowel constraints

Skills技術