Ph.D./M.S. Student at POSTECH NLP Lab Advised by Prof. Gary Geunbae Lee
Research Interests: Automatic Speech Recognition (ASR), Large Language Models (LLM), MLOps
2021.02 – Present
POSTECH (Pohang, Korea)
Ph.D./M.S. in Computer Science and Engineering
2017 – 2021
Sejong University (Seoul, Korea)
B.S. in Software Engineering & Data Science
Kanary 1B is a pre-release ~0.95B-parameter Korean ASR model built on NVIDIA’s canary-1b-v2, extending it with prompt-based control over punctuation, inverse text normalization, and foreign-word rendering tailored for Korean.
-
Grand Prize (2nd), Minister’s Award, Ministry of Science and ICT, 2022 link link
-
5th Place, Director’s Award, National Information Society Agency (NIA), 2023 link
-
4th Place, Director’s Award, National Information Society Agency (NIA), 2021 link
- Global Innovation Award, Minister of Science and ICT, 2025
Project: Multi-modal Multi-session Counseling System link link
- Best Paper Nominee, SIGDial 2024
“Enhancing Dialogue Speech Recognition with Robust Contextual Awareness via Noise Representation Learning” link
Speak & Spell: LLM-Driven Controllable Phonetic Error Augmentation for Robust Dialogue State Tracking
Jihyun Lee, Solee Im, Wonjun Lee, Gary Geunbae Lee
IJCNLP-AACL 2025, Mumbai, India link
DeRAGEC: Denoising Named Entity Candidates with Synthetic Rationale for ASR Error Correction
Wonjun Lee*, Solee Im*, Jinmyeong An, Yunsu Kim, Jungseul Ok, Gary Geunbae Lee
ACL Findings 2025, Vienna, Austria link
DyPCL: Dynamic Phoneme-level Contrastive Learning for Dysarthric Speech Recognition
Wonjun Lee*, Solee Im*, Heejin Do, Yunsu Kim, Jungseul Ok, Gary Geunbae Lee
NAACL 2025, Albuquerque, USA link
Enhancing Dialogue Speech Recognition with Robust Contextual Awareness via Noise Representation Learning
Wonjun Lee*, San Kim*, Gary Geunbae Lee
SIGDial 2024, Kyoto, Japan link
An Investigation Into Explainable Audio Hate Speech Detection
Wonjun Lee*, Jinmyeong An*, Yejin Jeon, Jungseul Ok, Yunsu Kim, Gary Geunbae Lee
SIGDial 2024, Kyoto, Japan link
Acoustic Feature Mixup for Balanced Multi-aspect Pronunciation Assessment
Heejin Do, Wonjun Lee, Gary Geunbae Lee
INTERSPEECH 2024, Kos Island, Greece link
Optimizing Two-Pass Cross-Lingual Transfer Learning: Phoneme Recognition and Phoneme to Grapheme Translation
Wonjun Lee, Yunsu Kim, Gary Geunbae Lee
ASRU 2023, Taipei, Taiwan link
Exploring the Viability of Synthetic Audio Data for Audio-Based Dialogue State Tracking
Jihyun Lee*, Yejin Jeon*, Wonjun Lee, Yunsu Kim, Gary Geunbae Lee
ASRU 2023, Taipei, Taiwan link
How to Use Speech Related Digital Biomarkers in Patients With Depressive Disorder
Wonjun Lee*, Seungyeon Seo*, Hyun Jeong Kim
Digital Health Research (Korean Society of Digital Health) link
한국어 자모단위 음성인식 결과 후보정을 위한 신경망 기반 자모 병합 방법론
임솔이*, 이원준*, 이근배, 김윤수
HCLT 2023, Jeju, Korea
다국어 음성인식을 위한 언어별 출력 계층 구조 Wav2Vec2.0
이원준, 이근배
HCLT 2021
Method and Apparatus for Multilingual Speech Recognition based on Artificial Intelligence Models
U.S. Patent, link