Newest papers

Weaponizing the Exit: Input Privacy Leakage via Exit Routing Manipulation in Multi-Exit Networks

Yulong Tian, Lei Wang, Jincan Yang, Hao Wu, Hao Han, Fengyuan Xu, Sheng Zhong · 2026 · arXiv

DOI: 10.48550/arXiv.2610.31246 · open-access PDF

Russian

Мы идентифицируем уязвимость конфиденциальности в облачных развертываниях этих сетей: вредоносный поставщик облачных услуг (CSP) может выводить чувствительные свойства отдельных входов пользователя исключительно из задержки вывода, не получая доступа к входам, промежуточным состояниям или выходам моделей. Атака использует контроль CSP над цепочкой поставок модели.

Machine translation (Meta M2M100); the English original follows.

Abstract

Multi-exit networks improve inference efficiency by allowing inputs to terminate early at intermediate exits. We identify a privacy vulnerability in cloud deployments of these networks: a malicious Cloud Service Provider (CSP) can infer sensitive properties of individual user inputs solely from inference latency, without accessing inputs, intermediate states, or model outputs. The attack exploits the CSP's control over the model supply chain. Prior to deployment, the CSP supplies a specially crafted multi-exit network manipulated so that inputs carrying a target sensitive property are routed t

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange