講演名 | 2022-07-28 Acceleration of HE-Transformer with bit reduced SEAL and HEXL Xinyi Li(Waseda Univ.), Masaki Nishi(Waseda Univ.), Teppei Shishido(Waseda Univ.), Keiji Kimura(Waseda Univ.), |
---|---|
PDFダウンロードページ | PDFダウンロードページへ |
抄録(和) | While cloud applications are covering people's privacy data, their related risks, such as theft risk, leak risk, and other serious risks, are becoming critical issues. Homomorphic encryption (HE) is a promising approach to overcome them since it enables calculation of encrypted data without decryption. Deep learning (DL) inference on cloud is an interesting application of HE. However, the computational cost of HE is too expensive, resulting in its limited usage in the real world. In this paper, we propose an acceleration technique of DL on HE, and implement it in HE-Transformer, which is a framework for DL on HE. Our approach is reducing the bit-width of internal data structure of HE libraries, SEAL and HEXL, from 64-bit to 32-bit. As a result, we can obtain up to $3.59times$ speedup compared to the original version, while still keeping data privacy. |
抄録(英) | While cloud applications are covering people's privacy data, their related risks, such as theft risk, leak risk, and other serious risks, are becoming critical issues. Homomorphic encryption (HE) is a promising approach to overcome them since it enables calculation of encrypted data without decryption. Deep learning (DL) inference on cloud is an interesting application of HE. However, the computational cost of HE is too expensive, resulting in its limited usage in the real world. In this paper, we propose an acceleration technique of DL on HE, and implement it in HE-Transformer, which is a framework for DL on HE. Our approach is reducing the bit-width of internal data structure of HE libraries, SEAL and HEXL, from 64-bit to 32-bit. As a result, we can obtain up to $3.59times$ speedup compared to the original version, while still keeping data privacy. |
キーワード(和) | Homomorphic Encryption / Deep Learning / SIMD |
キーワード(英) | Homomorphic Encryption / Deep Learning / SIMD |
資料番号 | CPSY2022-12,DC2022-12 |
発行日 | 2022-07-20 (CPSY, DC) |
研究会情報 | |
研究会 | CPSY / DC / IPSJ-ARC |
---|---|
開催期間 | 2022/7/27(から3日開催) |
開催地(和) | 海峡メッセ下関 |
開催地(英) | Kaikyo Messe Shimonoseki |
テーマ(和) | SWoPP2022: 並列/分散/協調システムとディペンダブルコンピューティングおよび一般 |
テーマ(英) | SWoPP2022: Parallel, Distributed and Cooperative Processing Systems and Dependable Computing |
委員長氏名(和) | 鯉渕 道紘(NII) / 土屋 達弘(阪大) / 津邑 公暁(名工大) |
委員長氏名(英) | Michihiro Koibuchi(NII) / Tatsuhiro Tsuchiya(Osaka Univ.) / 津邑 公暁(名工大) |
副委員長氏名(和) | 中島 耕太(富士通研) / 津邑 公暁(名工大) / 細川 利典(日大) |
副委員長氏名(英) | Kota Nakajima(Fujitsu Lab.) / Tomoaki Tsumura(Nagoya Inst. of Tech.) / Toshinori Hosokawa(Nihon Univ.) |
幹事氏名(和) | 井口 寧(北陸先端大) / 小川 周吾(日立) / 新井 雅之(日大) / 難波 一輝(千葉大) / 今村 智史(富士通研) / 谷本 輝夫(九大) / 新田 高庸(会津大) / 八巻 隼人(電通大) |
幹事氏名(英) | Yasushi Inoguchi(JAIST) / Shugo Ogawa(Hitachi) / Masayuki Arai(Nihon Univ.) / Kazuteru Namba(Chiba Univ.) / 今村 智史(富士通研) / 谷本 輝夫(九大) / 新田 高庸(会津大) / 八巻 隼人(電通大) |
幹事補佐氏名(和) | 小林 諒平(筑波大) / 宮島 敬明(明大) |
幹事補佐氏名(英) | Ryohei Kobayashi(Tsukuba Univ.) / Takaaki Miyajima(Meiji Univ.) |
講演論文情報詳細 | |
申込み研究会 | Technical Committee on Computer Systems / Technical Committee on Dependable Computing / Special Interest Group on System Architecture |
---|---|
本文の言語 | ENG |
タイトル(和) | |
サブタイトル(和) | |
タイトル(英) | Acceleration of HE-Transformer with bit reduced SEAL and HEXL |
サブタイトル(和) | |
キーワード(1)(和/英) | Homomorphic Encryption / Homomorphic Encryption |
キーワード(2)(和/英) | Deep Learning / Deep Learning |
キーワード(3)(和/英) | SIMD / SIMD |
第 1 著者 氏名(和/英) | Xinyi Li / Xinyi Li |
第 1 著者 所属(和/英) | Waseda University(略称:Waseda Univ.) Waseda University(略称:Waseda Univ.) |
第 2 著者 氏名(和/英) | Masaki Nishi / Masaki Nishi |
第 2 著者 所属(和/英) | Waseda University(略称:Waseda Univ.) Waseda University(略称:Waseda Univ.) |
第 3 著者 氏名(和/英) | Teppei Shishido / Teppei Shishido |
第 3 著者 所属(和/英) | Waseda University(略称:Waseda Univ.) Waseda University(略称:Waseda Univ.) |
第 4 著者 氏名(和/英) | Keiji Kimura / Keiji Kimura |
第 4 著者 所属(和/英) | Waseda University(略称:Waseda Univ.) Waseda University(略称:Waseda Univ.) |
発表年月日 | 2022-07-28 |
資料番号 | CPSY2022-12,DC2022-12 |
巻番号(vol) | vol.122 |
号番号(no) | CPSY-133,DC-134 |
ページ範囲 | pp.65-70(CPSY), pp.65-70(DC), |
ページ数 | 6 |
発行日 | 2022-07-20 (CPSY, DC) |