Low-Complexity DNN-Based End-To-End Automatic Speech Recognition using Low-Rank Approximation
- Title
- Low-Complexity DNN-Based End-To-End Automatic Speech Recognition using Low-Rank Approximation
- Authors
- Park, Jongmin; LEE, YOUNGJOO
- Date Issued
- 2020-10-23
- Publisher
- Institute of Electrical and Electronics Engineers Inc.
- Abstract
- Targeting the on-device speech-To-Text application for streaming inputs, this paper presents an efficient way to reduce the computational complexity of deep neural networks (DNNs) for attention-based speech processing. The proposed technique applies the singular value decomposition (SVD) to the large-sized matrix multiplications, removing less important computations by utilizing the low-rank approximation. The clipping thresholds are carefully adjusted to relax the computing costs as well as the memory overheads while maintaining the recognition accuracy.
- URI
- https://oasis.postech.ac.kr/handle/2014.oak/105839
- Article Type
- Conference
- Citation
- 17th International System-on-Chip Design Conference, ISOCC 2020, page. 210 - 211, 2020-10-23
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.