Presentation | 2024-03-22 Evaluation of Feature Inference Risk from Explainable AI metrics LIME and Shapley Values Ryotaro Toma, Hiroaki Kikuchi, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Explainability has gained attention to ensure fairness and transparency in machine learning models, providing users with a sense of understanding. Many services such as Amazon Web Services, Google Cloud Platform, and Microsoft Azure running Machine Learning as a Service (MLaaS) platforms, which provide several methods to explain model. However, in 2022, Luo et al. demonstrated that Shapley value-based explanations could lead to inference of private attribute, posing privacy risks of information leakage from models. Nevertheless, it remains unclear whether the attribute inference risk on the alternative explainability exist or not. Therefore, this study evaluates the attribute inference risk on LIME and compare the vulnerability with the explanability Shapley values. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | LIME / Shapley values / XAI / Explainability / Machine Learning / Feature Inference |
Paper # | ICSS2023-88 |
Date of Issue | 2024-03-14 (ICSS) |
Conference Information | |
Committee | ICSS / IPSJ-SPT |
---|---|
Conference Date | 2024/3/21(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | OIST |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Security, Trust, etc. |
Chair | Daisuke Inoue(NICT) |
Vice Chair | Akira Yamada(Kobe Univ.) / Toshihiro Yamauchi(Okayama Univ.) |
Secretary | Akira Yamada(Mitsubishi Electric) / Toshihiro Yamauchi(Univ. of Electro-Comm.) |
Assistant | Yo Kanemoto(NTT) / Masaya Sato(Okayama Prefectural 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) | Evaluation of Feature Inference Risk from Explainable AI metrics LIME and Shapley Values |
Sub Title (in English) | |
Keyword(1) | LIME |
Keyword(2) | Shapley values |
Keyword(3) | XAI |
Keyword(4) | Explainability |
Keyword(5) | Machine Learning |
Keyword(6) | Feature Inference |
1st Author's Name | Ryotaro Toma |
1st Author's Affiliation | Meiji University(Meiji Univ.) |
2nd Author's Name | Hiroaki Kikuchi |
2nd Author's Affiliation | Meiji University(Meiji Univ.) |
Date | 2024-03-22 |
Paper # | ICSS2023-88 |
Volume (vol) | vol.123 |
Number (no) | ICSS-448 |
Page | pp.pp.137-144(ICSS), |
#Pages | 8 |
Date of Issue | 2024-03-14 (ICSS) |