Chyi-Jiunn Lin

Chyi-Jiunn Lin

Research Assistant in Speech / NLP

NTU SPML Lab

About me

I received the Bachelor degree from National Taiwan University (NTU) in 2023, and currently work as a research assistant advised by Hung-yi Lee and Lin-shan Lee at Speech Processing and Machine Learning Laboratory, National Taiwan University. I am starting as a master’s student advised by Prof. Shinji Watanabe at Language Technologies Institute, Carnegie Mellon University in August 2024.

Interests
  • Spoken Language Understanding
  • Question Answering
  • Information Retrieval
  • Speech Synthesis
  • Speech Processing
  • Natural Language Processing
  • Machine Learning
Education
  • BS in Electrical Engineering, NTU

    GPA: 4.28/4.30 | Rank: 1/189

Experience

 
 
 
 
 
Speech Processing and Machine Learning Lab, NTU
Research Assistant
Speech Processing and Machine Learning Lab, NTU
August 2023 – Present Taipei

Research areas include:

  • Spoken Language Understanding
  • Question Answering
  • Information Retrieval
  • Speech Synthesis
  • Self-Supervised Learning Speech Representation
 
 
 
 
 
Mediatek Research
Deep Learning Research Intern
Mediatek Research
July 2021 – August 2021 Taipei
Worked on Dense Passage Retrieval with hierarchical representations for open-domain question answering.

All Publications

(2024). SpeechDPR: End-to-End Spoken Passage Retrieval for Open-Domain Spoken Question Answering. Accepted to IEEE ICASSP 2024.

PDF

(2023). SLUE Phase-2: A Benchmark Suite of Diverse Spoken Language Understanding Tasks. ACL 2023.

PDF Cite Code Dataset Project

(2023). Hierarchical Representations in Dense Passage Retrieval for Question-Answering. The FEVER workshop at EACL 2023.

PDF Cite Video

(2022). Meta-TTS: Meta-Learning for Few-Shot Speaker Adaptive Text-to-Speech. IEEE / ACM TASLP.

PDF Cite Code

Selected Projects

*
Spoken Language Understanding Evaluation (SLUE) Benchmark

SLUE provides:

  • Natural speech data for training and evaluation
  • Codes to pre-process datasets, train the baselines, and evaluate performance

for multiple SLU tasks.

Spoken Language Understanding Evaluation (SLUE) Benchmark
G.L.A.M.O.U.R.o.U.S in Micron UV Robot Design Contest
We developed an autonomous UV robot, G.L.A.M.O.U.R.o.U.S, for indoor disinfection. There I designed the navigational algorithm for the robot so that it could efficiently walk around the indoor environments without hitting obstacles. Finally we won 3rd place out of 64 teams worldwide.
G.L.A.M.O.U.R.o.U.S in Micron UV Robot Design Contest