Presentation | 2022-03-07 [Memorial Lecture] DistriHD: A Memory Efficient Distributed Binary Hyperdimensional Computing Architecture for Image Classification Dehua Liang, Jun Shiomi, Noriyuki Miura, Hiromitsu Awano, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Hyper-Dimensional (HD) computing is a brain-inspired learning approach for efficient and fast learning on today’s embedded devices. HD computing first encodes all data points to high-dimensional vectors called hypervectors and then efficiently performs the classification task using a well-defined set of operations. Although HD computing achieved reasonable performances in several practical tasks, it comes with huge memory requirements since the data point should be stored in a very long vector having thousands of bits. To alleviate this problem, we propose a novel HD computing architecture, called DistriHD which enables HD computing to be trained and tested using binary hypervectors and achieves high accuracy in single-pass training mode with significantly low hardware resources. DistriHD encodes data points to distributed binary hypervectors and eliminates the expensive item memory in the encoder, which significantly reduces the required hardware cost for inference. Our evaluation also shows that our model can achieve a 27.6× reduction in memory cost without hurting the classification accuracy. The hardware implementation also demonstrates that DistriHD achieves over 9.9× and 28.8× reduction in area and power, respectively. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Brain-inspired ComputingHyper-Dimensional ComputingMemory-EfficiencyDistributed System |
Paper # | VLD2021-84,HWS2021-61 |
Date of Issue | 2022-02-28 (VLD, HWS) |
Conference Information | |
Committee | VLD / HWS |
---|---|
Conference Date | 2022/3/7(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Online |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Design Technology for System-on-Silicon, Hardware Security, etc. |
Chair | Kazutoshi Kobayashi(Kyoto Inst. of Tech.) / Yasuhisa Shimazaki(Renesas Electronics) |
Vice Chair | Minako Ikeda(NTT) / Makoto Nagata(Kobe Univ.) / Daisuke Suzuki(Mitsubishi Electric) |
Secretary | Minako Ikeda(Osaka Univ.) / Makoto Nagata(NEC) / Daisuke Suzuki(NTT) |
Assistant |
Paper Information | |
Registration To | Technical Committee on VLSI Design Technologies / Technical Committee on Hardware Security |
---|---|
Language | ENG |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Memorial Lecture] DistriHD: A Memory Efficient Distributed Binary Hyperdimensional Computing Architecture for Image Classification |
Sub Title (in English) | |
Keyword(1) | Brain-inspired ComputingHyper-Dimensional ComputingMemory-EfficiencyDistributed System |
1st Author's Name | Dehua Liang |
1st Author's Affiliation | Osaka University(Osaka Univ.) |
2nd Author's Name | Jun Shiomi |
2nd Author's Affiliation | Osaka University(Osaka Univ.) |
3rd Author's Name | Noriyuki Miura |
3rd Author's Affiliation | Osaka University(Osaka Univ.) |
4th Author's Name | Hiromitsu Awano |
4th Author's Affiliation | Kyoto University(Kyoto Univ.) |
Date | 2022-03-07 |
Paper # | VLD2021-84,HWS2021-61 |
Volume (vol) | vol.121 |
Number (no) | VLD-412,HWS-413 |
Page | pp.pp.44-44(VLD), pp.44-44(HWS), |
#Pages | 1 |
Date of Issue | 2022-02-28 (VLD, HWS) |