Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory
With its high energy efficiency and ultra-high speed, processing-in-memory (PIM) technology is promising to enable high performance in Reservoir Computing (RC) systems. In this work, we demonstrate an RC system based on an as-fabricated PIM chip platform. The RC system extracts input into a high-dim...
| Main Authors: | , , , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/153571 |
| _version_ | 1826127020048777216 |
|---|---|
| author | Liu, Shuang Wu, Yuancong Xiong, Canlong Liu, Yihe Yang, Jing Yu, Q. Hu, S. G. Chen, Tupei Liu, Y. |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Liu, Shuang Wu, Yuancong Xiong, Canlong Liu, Yihe Yang, Jing Yu, Q. Hu, S. G. Chen, Tupei Liu, Y. |
| author_sort | Liu, Shuang |
| collection | NTU |
| description | With its high energy efficiency and ultra-high speed, processing-in-memory (PIM) technology is promising to enable high performance in Reservoir Computing (RC) systems. In this work, we demonstrate an RC system based on an as-fabricated PIM chip platform. The RC system extracts input into a high-dimensional space through the nonlinear characteristic and randomly connected reservoir states inside the PIM-based RC. To examine the system, nonlinear dynamic system predictions, including nonlinear auto-regressive moving average equation of order 10 driven time series, isolated spoken digit recognition task, and recognition of alphabet pronunciation, are carried out. The system saves about 50% energy and requires much fewer operations as compared with the RC system implemented with digital logic. This paves a pathway for the RC algorithm application in PIM with lower power consumption and less hardware resource required. |
| first_indexed | 2024-10-01T07:01:55Z |
| format | Journal Article |
| id | ntu-10356/153571 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:01:55Z |
| publishDate | 2021 |
| record_format | dspace |
| spelling | ntu-10356/1535712021-12-08T07:45:30Z Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory Liu, Shuang Wu, Yuancong Xiong, Canlong Liu, Yihe Yang, Jing Yu, Q. Hu, S. G. Chen, Tupei Liu, Y. School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Computation Theory Energy Efficiency With its high energy efficiency and ultra-high speed, processing-in-memory (PIM) technology is promising to enable high performance in Reservoir Computing (RC) systems. In this work, we demonstrate an RC system based on an as-fabricated PIM chip platform. The RC system extracts input into a high-dimensional space through the nonlinear characteristic and randomly connected reservoir states inside the PIM-based RC. To examine the system, nonlinear dynamic system predictions, including nonlinear auto-regressive moving average equation of order 10 driven time series, isolated spoken digit recognition task, and recognition of alphabet pronunciation, are carried out. The system saves about 50% energy and requires much fewer operations as compared with the RC system implemented with digital logic. This paves a pathway for the RC algorithm application in PIM with lower power consumption and less hardware resource required. Published version This work was supported by NSFC under Project Nos. 61774028, 92064004, and 61771097. 2021-12-08T06:29:33Z 2021-12-08T06:29:33Z 2021 Journal Article Liu, S., Wu, Y., Xiong, C., Liu, Y., Yang, J., Yu, Q., Hu, S. G., Chen, T. & Liu, Y. (2021). Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory. Applied Physics Letters, 119(10), 102103-. https://dx.doi.org/10.1063/5.0057132 0003-6951 https://hdl.handle.net/10356/153571 10.1063/5.0057132 2-s2.0-85114673635 10 119 102103 en Applied Physics Letters © 2021 Author(s). All rights reserved. This paper was published by AIP Publishing in Applied Physics Letters and is made available with permission of Author(s). application/pdf |
| spellingShingle | Engineering::Electrical and electronic engineering Computation Theory Energy Efficiency Liu, Shuang Wu, Yuancong Xiong, Canlong Liu, Yihe Yang, Jing Yu, Q. Hu, S. G. Chen, Tupei Liu, Y. Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title | Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title_full | Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title_fullStr | Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title_full_unstemmed | Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title_short | Efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing-in-memory |
| title_sort | efficient and reconfigurable reservoir computing to realize alphabet pronunciation recognition based on processing in memory |
| topic | Engineering::Electrical and electronic engineering Computation Theory Energy Efficiency |
| url | https://hdl.handle.net/10356/153571 |
| work_keys_str_mv | AT liushuang efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT wuyuancong efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT xiongcanlong efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT liuyihe efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT yangjing efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT yuq efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT husg efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT chentupei efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory AT liuy efficientandreconfigurablereservoircomputingtorealizealphabetpronunciationrecognitionbasedonprocessinginmemory |