A second-year Ph.D. student in National Taiwan University, advised by Prof. Hung-yi Lee. Research centers on efficiency, interpretability, and fine-tuning of large pre-trained speech models, with 7 first-author papers published at top-tier speech/NLP conferences or journals (IEEE TASLP, ACL, INTERSPEECH, IEEE ASRU, IEEE SLT) and several collaborative works. Recipient of multiple competitive scholarships and awards, including selection as an IEEE ASRU 2025 Best Student Paper Award candidate, the ISCA Travel Grant, GICE Elite Doctoral Scholarship, MOE Doctoral Scholarship, and CTCI Research Scholarship. Experienced in academic services as a reviewer for leading conferences. I am eager to explore new research areas and am currently seeking a full-time research position starting in 2027. If there are any opportunities for research collaboration, please feel free to contact me.
🔥 News
- 2026.06: 🎉🎉 Two papers accepted at INTERSPEECH 2026!
- 2026.04: 🎉🎉 Two paper accepted at ACL 2026 (1 main conference, 1 findings)! See you in San Diego 🇺🇸🏖️🐻❄️!
- 2025.11: 🎉🎉 Journal accepted at IEEE Transactions on Audio, Speech and Language Processing (TASLP)! Will be presented at ICASSP 2026. See you in Spain 🇪🇸!
- 2025.08: 🎉🎉 One paper accepted at APSIPA ASC 2025 main conference track. See you in Singapore 🇸🇬!
- 2025.08: 🎉🎉 One paper accepted at ASRU 2025 main conference track. See you in Hawaii 🇺🇸🥥🌴!
- 2024.09: 🎉🎉 Two papers accepted at SLT 2024 main conference track. See you in Macao 🇲🇴!
- 2024.06: 🎉🎉 Two papers accepted at Interspeech 2024. See you in Greece 🇬🇷!
- 2023.09: 🎉🎉 One paper accepeted at ASRU 2023.
📝 Selected Publications
An Exploration of Mamba for Speech Self-Supervised Models
Tzu-Quan Lin, Heng-Cheng Kuo, Tzu-Chieh Wei, Hsi-Chun Cheng, Chun-Wei Chen, Hsien-Fu Hsiao, Yu Tsao, Hung-yi Lee
ACL 2026 main conference
- This work explores Mamba-based HuBERT as a speech SSL model, showing its advantages in long-context and streaming ASR, improved speech unit quality, and competitive performance on probing tasks compared to Transformer-based models.
How Contrastive Decoding Enhances Large Audio Language Models?
Tzu-Quan Lin, Wei-Ping Huang, Yi-Cheng Lin, Hung-yi Lee
Under Review
- Developed a novel Transition Matrix framework to evaluate contrastive decoding in Large Audio Language Models, proving its efficacy in correcting audio blindness and uncertainty errors.
Tzu-Quan Lin, Wei-Ping Huang, Hao Tang, Hung-yi Lee
IEEE TASLP
- Speech-FT is a two-stage fine-tuning framework designed for speech representation learning. It improves performance on specific tasks while maintaining cross-task generalization ability.
- Speech-FT improves HuBERT’s performance on SUPERB by reducing phone error rate from 5.17% to 3.94%, lowering word error rate from 6.38% to 5.75%, and boosting speaker ID accuracy from 81.86% to 84.11%.
Is Smaller Always Faster? Tradeoffs in Compressing Self-Supervised Speech Transformers
Tzu-Quan Lin, Tsung-Huan Yang, Chun-Yao Chang, Kuang-Ming Chen, Tzu-hsun Feng, Hung-yi Lee, Hao Tang
ASRU 2025, Best Student Paper Finalist
- This work propose evaluating model compression methods using three different metrics: MACs, number of parameters, and real-time factor. We find that different compression methods excel in different metrics.
Identifying Speaker Information in Feed-Forward Layers of Self-Supervised Speech Transformers
Tzu-Quan Lin, Hsi-Chun Cheng, Hung-yi Lee, Hao Tang
APSIPA ASC 2025
- This work identifies speaker-relevant neurons in self-supervised speech Transformers and shows that preserving them during pruning helps maintain performance on speaker-related tasks.
Property Neurons in Self-Supervised Speech Transformers
Tzu-Quan Lin, Guan-Ting Lin, Hung-yi Lee, Hao Tang
IEEE SLT 2024
- In this work, we identify a set of property neurons in the feedforward layers of Transformers to study how speech-related properties, such as phones, gender, and pitch, are stored.
DAISY: Data Adaptive Self-Supervised Early Exit for Speech Representation Models
Tzu-Quan Lin, Hung-yi Lee, Hao Tang
Interspeech 2024
- This work introduces a novel early exit method for speech self-supervised models that enhances the speed of HuBERT with minimal performance loss.
MelHuBERT: A Simplified Hubert on Mel Spectrograms
Tzu-Quan Lin, Hung-yi Lee, Hao Tang
IEEE ASRU 2023
- MelHuBERT simplifies the model architecture and loss function of HuBERT, achieving comparable performance while saving 33.5% of MACs per one second of speech.
- CAAD: Contrastive Audio-Aware Distillation for Efficient Speech Language Models, Chun-Wei Chen, Tzu-Quan Lin, Ke-Han Lu, Wei-Ping Huang, Hung-Yi Lee. INTERSPEECH 2026
- AdaLTM: Adaptive Layer-wise Task Vector Merging for Categorical Speech Emotion Recognition with ASR Knowledge Integration, Chia-Yu Lee, Huang-Cheng Chou*, Tzu-Quan Lin*, Yuanchao Li*, Ya-Tse Wu, Shrikanth Narayanan, Chi-Chun Lee. INTERSPEECH 2026
- Listen and Speak Fairly: A Study on Semantic Gender Bias in Speech Integrated Large Language Models, Yi-Cheng Lin, Tzu-Quan Lin, Chih-Kai Yang, Ke-Han Lu, Wei-Chih Chen, Chun-Yi Kuan, Hung-yi Lee. IEEE SLT 2024
- On the social bias of speech self-supervised models, Yi-Cheng Lin, Tzu-Quan Lin, Hsi-Che Lin, Andy t. Liu, Hung-yi Lee. Interspeech 2024, Best Paper Runner-up in Special Session
- Superb @ SLT 2022: Challenge on Generalization and Efficiency of Self-Supervised Speech Representation Learning, Tzu-hsun Feng, Annie Dong, Ching-Feng Yeh, Shu-wen Yang, Tzu-Quan Lin, Jiatong Shi, et al.. IEEE SLT 2022, Best Paper Finalists
- Pseudo2Real: Task Arithmetic for Pseudo-Label Correction in Automatic Speech Recognition, Yi-Cheng Lin, Yu-Hsuan Li Liang, Hsuan Su, Tzu-Quan Lin, Shang-Tse Chen, Yun-Nung Chen, Hung-yi Lee, ACL 2026 Findings
📖 Educations
- 2022.07 - now, PhD in Electrical, Electronics, Communications Engineering (EE), Data Science and Smart Networking, National Taiwan University (transferred from M.S. program in 2024.09)
- 2018.09 - 2022.06, Bachelor in Department of Computer Science and Information Engineering (CSIE), National Taiwan University
🏆 Honors and Awards
- 2025, IEEE ASRU 2025 Best Student Paper Finalist, selected as one of 11 out of 215 accepted papers
- 2025, CTCI Research Scholarship
- 2024, Best Paper Runner-up, INTERSPEECH 2024 Special Session
- 2024, Interspeech 2024 Travel Grant
- 2024, GICE Elite Doctoral Scholarship
- 2024, MOE Doctoral Scholarship
- 2022, Best Paper Finalist, IEEE SLT 2022, 6 out of 363 accepted papers
📚 Academic Services
- 2026, Reviewer for NeurIPS 2026, EMNLP 2026, IEEE TASLP, INTERSPEECH 2026, SLT 2026
- 2025, Reviewer for EMNLP 2025, EACL 2026, ICASSP 2026, ACML 2025, ROCLING 2025
- 2024, Reviewer for NeurIPS 2024, ISCSLP 2024
🔬 Research Experience
- 2026.01 - 2026.06, Deep Learning Research Intern, MediaTek Research, Taipei, Taiwan.
- Revisited the narrow-wide-narrow Transformer FFN convention with residual hourglass bottlenecks, achieving comparable language-modeling quality up to 8B parameters with 2-4% fewer training FLOPs and up to 1.9x faster 64k-context decoding.
- 2022.09 - now, Graduate Researcher, Speech Processing and Machine Learning Lab, National Taiwan University.
- Led multiple projects with internal and external collaborators on efficiency, interpretability, and fine-tuning of pre-trained speech models, publishing 7 first-author papers at top-tier speech/NLP conferences and journals.
- 2022.07, Research Team Leader, 8th JSALT Summer Workshop, Virtual (hosted from Baltimore, USA).
- Studied how to simplify pre-trained speech models while maintaining comparable performance, and highlighted key tradeoffs between model size, speed, and performance in compression methods.
- 2021.07 - 2021-09, Machine Learning Engineer Intern, aetherAI, Taipei, Taiwan.
- Researched a novel method for X-ray image fracture detection, focusing on both localization and severity classification.