معرفی
Hao Tang is a Lecturer in Speech Technology at the School of Informatics, University of Edinburgh, affiliated with the Institute for Language, Cognition and Computation (ILCC) and the Centre for Speech Technology Research (CSTR). He contributes to cutting-edge research in speech and language processing, with a focus on speech representations and self-supervised learning.
- PhD, Toyota Technological Institute at Chicago (2017–2020), advised by Karen Livescu
- Postdoctoral Associate, MIT Spoken Language Systems Group (2020–)
- Master’s, National Taiwan University, advised by Lin-Shan Lee
Hao Tang’s research centers on speech representations, particularly discrete and geometric properties in self-supervised models. He investigates how speech systems encode speaker and phonetic information, and how these representations can be improved for tasks like phone segmentation, acoustic word embedding, and text-to-speech. He also works on text summarization, especially opinion and attributable summarization. His work bridges machine learning, cognitive modeling, and practical speech applications.
His recent publications span top venues including Interspeech, ICASSP, ACL, NeurIPS, and IEEE/ACM Transactions. Key themes include self-supervised learning, disentangled representations, predictive coding, and efficient speech modeling. He has co-authored papers on discrete speech units, orthogonality in representations, and context modeling in neural speech systems.
- Best Student Paper Award, Interspeech 2020
- Computational Modeling Prize for Perception & Action, CogSci 2024
- Speech and Language Processing Student Paper Award, ICASSP 2016
- Best Student Paper of Speech and Language Processing, ICASSP 2016
- Best Paper Nominee, ASRU 2015
Hao Tang actively supervises PhD and Master’s students, many of whom have published at leading conferences and gone on to positions at Cohere, Apple, Sesame, and top PhD programs. He serves as Associate Member of IEEE SLTC, Meta Reviewer for ICASSP, Area Chair for ACL ARR, Action Editor for TACL, and Area Chair for ICLR and NeurIPS. He teaches core courses such as Machine Learning (INFR10086) and Automatic Speech Recognition (INFR11033).
Hao Tang leads a research group within ILCC and CSTR, collaborating with researchers like Sharon Goldwater, Jim Glass, and Karen Livescu. His team focuses on developing interpretable, efficient, and scalable models for speech and language understanding.
