| Presentation | 2021-03-02 Research on the vulnerability of homoglyph attacks to online machine translation system Takeshi Sakamoto, Tatsuya Mori, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | It has been widely known that systems empowered by neural network algorithms are vulnerable against an intrinsic attack named ``Adversarial Input'', which can be generated by adding small perturbations to the original inputs, aiming at fooling the systems. Adversarial Input has the characteristic that the difference from the original input is not recognized by humans. In this research, we investigate the output of eight machine translation systems that can be used online when homoglyphs and special characters are input as malicious inputs, and evaluate their vulnerability to hostile attacks. As a result of the investigation, it was found that it is possible to estimate the preprocessing to each machine translation system. In this paper, we propose vulnerabilities and countermeasures specific to each system based on the estimation results of preprocessing. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Adversarial Example / Neural Machine Translation / Preprocess / Homoglyph |
| Paper # | ICSS2020-50 |
| Date of Issue | 2021-02-22 (ICSS) |
| Conference Information | |
| Committee | ICSS / IPSJ-SPT |
|---|---|
| Conference Date | 2021/3/1(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Security, Trust, etc. |
| Chair | Hiroki Takakura(NII) |
| Vice Chair | Katsunari Yoshioka(Yokohama National Univ.) / Kazunori Kamiya(NTT) |
| Secretary | Katsunari Yoshioka(NICT) / Kazunori Kamiya(KDDI labs.) |
| Assistant | Keisuke Kito(Mitsubishi Electric) / Toshihiro Yamauchi(Okayama Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Information and Communication System Security / Special Interest Group on Security Psychology and Trust |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Research on the vulnerability of homoglyph attacks to online machine translation system |
| Sub Title (in English) | |
| Keyword(1) | Adversarial Example |
| Keyword(2) | Neural Machine Translation |
| Keyword(3) | Preprocess |
| Keyword(4) | Homoglyph |
| 1st Author's Name | Takeshi Sakamoto |
| 1st Author's Affiliation | Waseda University(Waseda Univ) |
| 2nd Author's Name | Tatsuya Mori |
| 2nd Author's Affiliation | Waseda University(Waseda Univ) |
| Date | 2021-03-02 |
| Paper # | ICSS2020-50 |
| Volume (vol) | vol.120 |
| Number (no) | ICSS-384 |
| Page | pp.pp.144-149(ICSS), |
| #Pages | 6 |
| Date of Issue | 2021-02-22 (ICSS) |