Cited time in webofscience Cited time in scopus

Approximate LSTM Computing for Energy-Efficient Speech Recognition

Title
Approximate LSTM Computing for Energy-Efficient Speech Recognition
Author(s)
Jo, JunseoKung, JaehaLee, Youngjoo
Issued Date
2020-12
Citation
Electronics, v.9, no.12, pp.2004 - 13
Type
Article
Author Keywords
approximate computingLSTM architecturespeech recognitionVLSI design
ISSN
2079-9292
Abstract
This paper presents an approximate computing method of long short-term memory (LSTM) operations for energy-efficient end-to-end speech recognition. We newly introduce the concept of similarity score, which can measure how much the inputs of two adjacent LSTM cells are similar to each other. Then, we disable the highly-similar LSTM operations and directly transfer the prior results for reducing the computational costs of speech recognition. The pseudo-LSTM operation is additionally defined for providing the approximate computation with reduced processing resolution, which can further relax the processing overheads without degrading the accuracy. In order to verify the proposed idea, in addition, we design an approximate LSTM accelerator in 65 nm CMOS process. The proposed accelerator newly utilizes a number of approximate processing elements (PEs) to support the proposed skipped-LSTM and pseudo-LSTM operations without degrading the energy efficiency. Moreover, sparsity-aware scheduling is introduced by introducing the small-sized on-chip SRAM buffer. As a result, the proposed work provides an energy-efficient but still accurate speech recognition system, which consumes 2.19 times less energy than the baseline architecture. © 2020 by the authors. Licensee MDPI, Basel, Switzerland.
URI
http://hdl.handle.net/20.500.11750/12555
DOI
10.3390/electronics9122004
Publisher
MDPI AG
Related Researcher
  • 궁재하 Kung, Jaeha
  • Research Interests 딥러닝; 가속하드웨어; 저전력 하드웨어; 고성능 시스템
Files in This Item:
2_s2.0_85096839190.pdf

2_s2.0_85096839190.pdf

기타 데이터 / 3.12 MB / Adobe PDF download
Appears in Collections:
Department of Electrical Engineering and Computer Science Intelligent Digital Systems Lab 1. Journal Articles

qrcode

  • twitter
  • facebook
  • mendeley

Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.

BROWSE