| Presentation | 2021-07-21 Lossy Error Correction Coding for Vector-Matrix Multiplication Leo Otani, Haruhiko Kaneko, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Error control codes are effective to improve the reliability and precision of the vector-matrix multiplications executed on unreliable low-precision hardware. Since numerical errors of small magnitude may not affect the reliability or performance of system significantly in some applications, such as neural networks and image processings, this report presents error control codes which can efficiently correct large magnitude errors while allowing small magnitude errors. The proposed coding modifies the systematic encoding of nonbinary polar code, that is, we improve the code rate by diverting some check symbols to information symbols, while allowing small magnitude errors in the information symbols. Simulation results show that the proposed coding has higher random error correction capability compared to existing method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | ReRAM crossbar array / vector-matrix multiplication / polar code / successive cancellation decoding |
| Paper # | CPSY2021-7,DC2021-7 |
| Date of Issue | 2021-07-13 (CPSY, DC) |
| Conference Information | |
| Committee | CPSY / DC / IPSJ-ARC |
|---|---|
| Conference Date | 2021/7/20(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | SWoPP2021: Parallel, Distributed and Cooperative Processing Systems and Dependable Computing |
| Chair | Michihiro Koibuchi(NII) / Hiroshi Takahashi(Ehime Univ.) / Hiroshi Inoue(Kyushu Univ.) |
| Vice Chair | Kota Nakajima(Fujitsu Lab.) / Tomoaki Tsumura(Nagoya Inst. of Tech.) / Tatsuhiro Tsuchiya(Osaka Univ.) |
| Secretary | Kota Nakajima(JAIST) / Tomoaki Tsumura(Hitachi) / Tatsuhiro Tsuchiya(Nihon Univ.) / (Chiba Univ.) |
| Assistant | Ryohei Kobayashi(Tsukuba Univ.) / Takaaki Miyajima(Meiji Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Computer Systems / Technical Committee on Dependable Computing / Special Interest Group on System Architecture |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Lossy Error Correction Coding for Vector-Matrix Multiplication |
| Sub Title (in English) | |
| Keyword(1) | ReRAM crossbar array |
| Keyword(2) | vector-matrix multiplication |
| Keyword(3) | polar code |
| Keyword(4) | successive cancellation decoding |
| 1st Author's Name | Leo Otani |
| 1st Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| 2nd Author's Name | Haruhiko Kaneko |
| 2nd Author's Affiliation | Tokyo Institute of Technology(Tokyo Tech) |
| Date | 2021-07-21 |
| Paper # | CPSY2021-7,DC2021-7 |
| Volume (vol) | vol.121 |
| Number (no) | CPSY-116,DC-117 |
| Page | pp.pp.37-42(CPSY), pp.37-42(DC), |
| #Pages | 6 |
| Date of Issue | 2021-07-13 (CPSY, DC) |