My name is Zezhong Jin, and I am currently a Ph.D. student in the Department of Electrical and Electronic Engineering at The Hong Kong Polytechnic University. My research interests include:

• Robust Speaker Representation Learning
• Generative Models
• Speech Large Language Models
• Automatic Speech Recognition

[Google Scholar]   [Email]

🔥 News

• 2026: Tutorial speaker for Speech Large Language Models: Architectures, Efficient Adaptation, and Applications at ICME 2026.
• 2026: Two papers were accepted by INTERSPEECH 2026. See you in Australia!
• 2026: Distilling Attention Knowledge for Speaker Verification was accepted by ICASSP 2026.
• 2025: Uncertainty-Aware Multi-Head Multi-Mode Knowledge Distillation for Self-Supervised Speaker Verification was accepted by IEEE/ACM TASLP.
• 2025: TrInk: Ink Generation with Transformer Network was accepted by EMNLP 2025.
• 2025: Denoising Student Features with Diffusion Models for Knowledge Distillation in Speaker Verification was accepted by ICASSP 2025.

📝 Publications

• EMNLP 2025TrInk: Ink Generation with Transformer Network, Zezhong Jin, S. Desai, X. Chen, B. Fang, Z. Huang, Z. Li, C.-X. Gan, X. Tu, M.-W. Mak, Y. Lu, and S. Liu.
• TASLPUncertainty-Aware Multi-Head Multi-Mode Knowledge Distillation for Self-Supervised Speaker Verification, Zezhong Jin, Y. Tu, Z. Li, M.-W. Mak, and K.-A. Lee.
• ICASSP 2025Denoising Student Features with Diffusion Models for Knowledge Distillation in Speaker Verification, Zezhong Jin, Y. Tu, Z. Li, Z. Huang, C.-X. Gan, and M.-W. Mak.
• NeurocomputingAdversarially Adaptive Temperatures for Decoupled Knowledge Distillation with Applications to Speaker Verification, Zezhong Jin, Y. Tu, C.-X. Gan, and M.-W. Mak.
• APSIPA ASC 2024JOSEPH: Phonetic-Aware Speaker Embedding for Far-Field Speaker Verification, Zezhong Jin, Y. Tu, and M.-W. Mak.
• INTERSPEECH 2024Self-Supervised Learning with Multi-Head Multi-Mode Knowledge Distillation for Speaker Verification, Zezhong Jin, Y. Tu, and M.-W. Mak.
• INTERSPEECH 2024W-GVKT: Within-Global-View Knowledge Transfer for Speaker Verification, Zezhong Jin, Y. Tu, and M.-W. Mak.
• ICASSP 2026Distilling Attention Knowledge for Speaker Verification, Zezhong Jin, S. Liu, Z. Li, K.-A. Lee, and M.-W. Mak.
• INTERSPEECH 2026Beyond Residual Connections: Manifold-Constrained Hyper-Connections for Robust Speaker Representation Learning, Zezhong Jin, X. Wang, Z. Li, M.-W. Mak, and K.-A. Lee.
• INTERSPEECH 2025IDIR: Identifying and Distilling Informative Relations for Speaker Verification, C.-X. Gan, Z. Li, Zezhong Jin, Z. Huang, M.-W. Mak, and K.-A. Lee.
• INTERSPEECH 2025Disentangling Speaker and Content in Pre-trained Speech Models with Latent Diffusion for Robust Speaker Verification, Z. Li, M.-W. Mak, J.-T. Chien, M. Pilanci, Zezhong Jin, and H. Meng.
• TASLPDisentangling Speech Representations Learning with Latent Diffusion for Speaker Verification, Z. Li, M.-W. Mak, J.-T. Chien, M. Pilanci, Zezhong Jin, and H. Meng.
• ICASSP 2025Grouped Knowledge Distillation with Adaptive Logit Softening for Speaker Recognition, C.-X. Gan, Y. Tu, Zezhong Jin, M.-W. Mak, and K.-A. Lee.
• INTERSPEECH 2026DualSCL: Emotion Recognition in Conversation through Multimodal Alignment and Emotional Inertia Modeling, Z. Huang, M.-W. Mak, K.-A. Lee, C.-X. Gan, Zezhong Jin, and R. Zuo.

📖 Education

• May 2023 - Present, The Hong Kong Polytechnic University
  Ph.D. in Electrical and Electronic Engineering, Electronic Information (Artificial Intelligence).

• August 2021 - May 2023, The Hong Kong Polytechnic University
  M.S. in Electrical and Electronic Engineering, Electronic Information (Artificial Intelligence).

• September 2017 - June 2021, Hebei University
  B.Eng. in Electronic Information, Software Engineering Track.

💻 Internships

• Research Intern, ByteDance, China - ‘Jindouyun’ (筋斗云) Talent Program (June 2026 - Present).
• Research Intern, Tencent, China (January 2026 - May 2026).
• Research Intern, Microsoft Research Asia, China (November 2024 - November 2025). Co-supervised by Dr. Shujie Liu and Dr. Jinyu Li.
• Research Intern, Zoom, China (April 2022 - August 2022).

🧑‍🔬 Services

Reviewer:

• ICASSP 2024, 2025, 2026
• INTERSPEECH 2025
• APSIPA 2025

🎖 Honors and Awards

• Outstanding Graduate, The Hong Kong Polytechnic University, 2023.
• Outstanding Thesis Award, The Hong Kong Polytechnic University, 2023.
• The Hong Kong Polytechnic University Scholarship, 2022.
• First-Class Scholarship, Hebei University, 2018, 2019, 2020.

📚 Patents

• ONLINE INK GENERATION USING A TRANSFORMER NETWORK, Invention Patent, No. 505283-US01.
• Electronic Information Protection Device, Utility Model Patent, Patent No. ZL 2019 2 0216847.1.

💬 Talks

• June 2024, Self-Supervised Learning with Multi-Head Multi-Mode Knowledge Distillation for Speaker Verification, INTERSPEECH 2024 paper preview seminar.
• June 2024, W-GVKT: Within-Global-View Knowledge Transfer for Speaker Verification, INTERSPEECH 2024 paper preview seminar.
• March 2025, Denoising Student Features with Diffusion Models for Knowledge Distillation in Speaker Verification, ICASSP 2025 paper preview seminar.
• April 2026, Distilling Attention Knowledge for Speaker Verification, ICASSP 2026 paper preview seminar.
• July 2026, Speech Large Language Models: Architectures, Efficient Adaptation, and Applications, ICME 2026 Tutorial.