Hao Tang
Lecturer in Speech Technology
Studies speech and language processing, particularly self-supervised speech representations and models of who says what, when and where.
Hao Tang's lab at Edinburgh works on speech recognition & spoken language and generative & foundation models. 29 papers since 2023, 10 at top venues, mostly Interspeech and SLT. 7 PhD students. 8 alumni, most often to Apple and Cerence AI.
At a glance
- papers, 2023–2026
- 8/ year
- at T1 venues
- 34%
- PhD students
- 7
- MS students
- 0
- alumni listed
- 8
Papers by yearfrom the lab homepage, 2023–2026
- 6 papers2026 · T1 2 · collecting
- 6 papers2025 · T1 1
- 7 papers2024 · T1 2
- 10 papers2023 · T1 5
T1 venuesOtherVenues: Interspeech, SLT, ICASSP, ASRU, CoNLL
Research focuspaper titles that use each topic's words
- Speech recognition & spoken language19 papers
- speech19 papers
- Generative & foundation models9 papers
- self-supervised speech5 papers
- studies speech0 papers
Paper list
From the lab homepage.
20266 papers
- CoNLLA framework for analyzing concept representations in neural models (opens in a new tab)
- SLTDeep speech representations track word timing along a low-dimensional circular trajectory (opens in a new tab)
- ICASSP (T1)Identifying the minimal and maximal phonetic subspace of speech representations (opens in a new tab)
- SLTIn-context adaptation of encoder-decoder models in speech recognition
- Transactions of the ACLLearning speech representations with variational predictive coding (opens in a new tab)
- EMNLP (T1)MELD: Mel-spectrogram-based speech language modeling with discrete latent variables (opens in a new tab)
20256 papers
- Interspeech (T1)Effective context in neural speech models (opens in a new tab)
- Transactions of the ACLEnd-to-end long document summarization using gradient caching (opens in a new tab)
- APSIPAIdentifying speaker information in feed-forward layers of self-supervised speech Transformers (opens in a new tab)
- ASRUIs smaller always faster? Tradeoffs in compressing self-supervised speech Transformers (opens in a new tab)
- —Speech-FT: Merging pre-trained and fine-tuned speech representation models for cross-task generalization
- ASRUWhisper has an internal word aligner (opens in a new tab)
20247 papers
- CogSciA predictive learning model can simulate temporal dynamics and context effects found in neural representations of continuous speech (opens in a new tab)
- SLTA simple HMM with self-supervised representations for phone segmentation (opens in a new tab)
- Interspeech (T1)DAISY: Data adaptive self-supervised early exit for speech representation models (opens in a new tab)
- SLTEstimating the completeness of discrete speech units (opens in a new tab)
- Transactions of the ACLHierarchical indexing for retrieval-augmented opinion summarization (opens in a new tab)
- Interspeech (T1)Orthogonality and isotropy of speaker and phonetic information in self-supervised speech representations
- SLTProperty neurons in self-supervised speech Transformers (opens in a new tab)
202310 papers
- —ACL, 2023
- Interspeech (T1)Acoustic word embeddings for untranscribed target languages with continued pretraining and learned pooling (opens in a new tab)
- ICASSP (T1)Analyzing acoustic word embeddings from pre-trained self-supervised speech models (opens in a new tab)
- —ASRU, 2023
- ICASSP (T1)Conditioning and sampling in variational diffusion models for speech super-resolution (opens in a new tab)
- —Improving Seq2Seq TTS frontends with transcribed speech audio (opens in a new tab)
- ICASSP (T1)Learning dependencies of discrete speech representations with neural hidden Markov models (opens in a new tab)
- ASRUMelHuBERT: A simplified HuBERT on Mel spectrograms (opens in a new tab)
- Interspeech (T1)Self-supervised predictive coding models encode speaker and phonetic information in orthogonal subspaces (opens in a new tab)
- SSWSpell4TTS: Acoustically-informed spellings for improving text-to-speech pronunciations (opens in a new tab)
Where alumni went8 listed, 67% to industry
Industry 4Academia 2
Most often to
- Apple
- Cerence AI
- Cohere
Also to
- PhD at University of Auckland
- PhD at University of Sheffield
Apply to EdinburghFall 2027
Not announced yet
Informatics PhD
- Round
- Rolling (funding consideration)
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- none
Not yet confirmed for this intake.
Similar labsnearest by research summary
- GBEdinburgh
Mirella Lapata
- Multilingual, translation & linguistics
- CNZhejiang
Ming Cai
0 T1 - CNTsinghua
Lijie Wen
25 T1 - CAToronto
Barend Beekhuizen
- Multilingual, translation & linguistics
- GBOxford
Janet B. Pierrehumbert
- Large language models
- USUSC
Jonathan May
- Multilingual, translation & linguistics
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.