Omni-Sparsity DNN: Fast Sparsity Optimization for On-Device Streaming E2E ASR Via Supernet-Reference-Cited by-同舟云学术

Omni-Sparsity DNN: Fast Sparsity Optimization for On-Device Streaming E2E ASR Via Supernet

Published:2022-05-23 Issue: Volume: Page:
ISSN:
Container-title:ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
language:
Short-container-title:

Author:

Yang Haichuan¹,Shangguan Yuan¹,Wang Dilin¹,Li Meng¹,Chuang Pierce¹,Zhang Xiaohui¹,Venkatesh Ganesh¹,Kalinli Ozlem¹,Chandra Vikas¹

Affiliation:

1. Facebook AI

Publisher

IEEE

Link

Reference22 articles.

1. Sequence transduction with recurrent neural networks;graves;Proc of ICML Workshop on Representation Learning,2012

3. Slimmable neural networks;yu;Proc of LCLR,2018

4. Alphanet: Improved training of super-net with alpha-divergence;wang;Proc of ICML,2021

5. Optimizing speech recognition for the edge;shangguan;On-device Intelligence Workshop at MLSys,2019

Cited by 7 articles. 订阅此论文施引文献订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献

1. TODM: Train Once Deploy Many Efficient Supernet-Based RNN-T Compression For On-Device ASR Models;ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2024-04-14

2. Dynamic ASR Pathways: An Adaptive Masking Approach Towards Efficient Pruning of a Multilingual ASR Model;ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2024-04-14

3. Small-Footprint Slimmable Networks for Keyword Spotting;ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2023-06-04

4. Continual Learning for On-Device Speech Recognition Using Disentangled Conformers;ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2023-06-04

5. Learning ASR Pathways: A Sparse Multilingual ASR Model;ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP);2023-06-04