1 code implementation • 9 Oct 2022 • Huanru Henry Mao
Autoregressive Transformers are strong language models but incur O(T) complexity during per-token generation due to the self-attention mechanism.
no code implementations • 1 Jul 2020 • Huanru Henry Mao
It involves first pre-training a model on a large amount of unlabeled data, then adapting the model to target tasks of interest.
3 code implementations • 16 May 2020 • Huanru Henry Mao, Shuyang Li, Julian McAuley, Garrison Cottrell
Speech recognition (ASR) and speaker diarization (SD) models have traditionally been trained separately to produce rich conversation transcripts with speaker labels.
13 code implementations • 10 Mar 2020 • Thomas Bachlechner, Bodhisattwa Prasad Majumder, Huanru Henry Mao, Garrison W. Cottrell, Julian McAuley
Deep networks often suffer from vanishing or exploding gradients due to inefficient signal propagation, leading to long training times or convergence difficulties.
1 code implementation • IJCNLP 2019 • Huanru Henry Mao, Bodhisattwa Prasad Majumder, Julian McAuley, Garrison W. Cottrell
Stories generated with neural language models have shown promise in grammatical and stylistic consistency.
1 code implementation • 10 Jul 2019 • Chris Donahue, Huanru Henry Mao, Yiting Ethan Li, Garrison W. Cottrell, Julian McAuley
We are interested in the task of generating multi-instrumental music scores.
2 code implementations • 12 Jun 2018 • Chris Donahue, Huanru Henry Mao, Julian McAuley
Existing research on music generation focuses on composition, but often ignores the expressive performance characteristics required for plausible renditions of resultant pieces.
1 code implementation • 3 Jan 2018 • Huanru Henry Mao, Taylor Shin, Garrison W. Cottrell
Recent advances in deep neural networks have enabled algorithms to compose music that is comparable to music composed by humans.
Sound Audio and Speech Processing