Enhanced membership inference attack method leveraging data characteristics

Aiming at the problem that the existing membership inference attack (MIA) methods only present generic testing results and had not sufficiently evaluated the effectiveness of such attacks, which could mislead attackers, multiple test scenarios were first constructed and various evaluation metrics we...

Descrizione completa

Dettagli Bibliografici

Autori principali:	NIU Jun, SHEN Kuo, WANG Yuecong, HAN Xuexue, ZHANG Jiawei, LI Xinghua, ZHANG Yuqing
Natura:	Artigo
Lingua:	Chinês
Pubblicazione:	Editorial Department of Journal on Communications 2025-09-01
Serie:	Tongxin xuebao
Soggetti:	machine learning membership inference attack data security data characteristics
Accesso online:	http://www.joconline.com.cn/zh/article/doi/10.11959/j.issn.1000-436x.2025162/
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!