Related papers: LIPSTICK: Corruptibility-Aware and Explainable Gra…

Due to significant improvements in performance in recent years, neural networks are currently used for an ever-increasing number of applications. However, neural networks have the drawback that their decisions are not readily interpretable…

The pi calculus is a basic theory of mobile communication based on the notion of interaction, which, aimed at analyzing and modelling the behaviors of communication process in communicating and mobile systems, is widely applied to the…

Mitigating reward hacking--where AI systems misbehave due to flaws or misspecifications in their learning objectives--remains a key challenge in constructing capable and aligned models. We show that we can monitor a frontier reasoning…

Graph neural networks (GNNs) are widely used for learning from graph-structured data in domains such as social networks, recommender systems, and financial platforms. To comply with privacy regulations like the GDPR, CCPA, and PIPEDA,…

This paper describes the design, implementation, and evaluation of Otak, a system that allows two non-colluding cloud providers to run machine learning (ML) inference without knowing the inputs to inference. Prior work for this problem…

Knowledge erasure in large language models (LLMs) is important for ensuring compliance with data and AI regulations, safeguarding user privacy, mitigating bias, and misinformation. Existing unlearning methods aim to make the process of…

Reinforcement learning with verifiable rewards (RLVR) typically optimizes for outcome rewards without imposing constraints on intermediate reasoning. This leaves training susceptible to reward hacking, where models exploit loopholes (e.g.,…

Safety alignment is critical for the ethical deployment of large language models (LLMs), guiding them to avoid generating harmful or unethical content. Current alignment techniques, such as supervised fine-tuning and reinforcement learning…

Due to its sound theoretical basis and practical efficiency, masking has become the most prominent countermeasure to protect cryptographic implementations against physical side-channel attacks (SCAs). The core idea of masking is to randomly…

A popular class of defenses against prompt injection attacks on large language models (LLMs) relies on fine-tuning to separate instructions and data, so that the LLM does not follow instructions that might be present with data. We evaluate…

The participation of third-party entities in the globalized semiconductor supply chain introduces potential security vulnerabilities, such as intellectual property piracy and hardware Trojan (HT) insertion. Graph neural networks (GNNs) have…

Timing-based side and covert channels in processor caches continue to be a threat to modern computers. This work shows for the first time a systematic, large-scale analysis of Arm devices and the detailed results of attacks the processors…

As LLMs advance into autonomous agents with tool-use capabilities, they introduce security challenges that extend beyond traditional content-based LLM safety concerns. This paper introduces Sequential Tool Attack Chaining (STAC), a novel…

Federated Learning (FL) is a distributed learning paradigm that enables different parties to train a model together for high quality and strong privacy protection. In this scenario, individual participants may get compromised and perform…

Graph neural networks (GNNs) have exhibited superior performance in various classification tasks on graph-structured data. However, they encounter the potential vulnerability from the link stealing attacks, which can infer the presence of a…

Machine learning models deployed as a service (MLaaS) are susceptible to model stealing attacks, where an adversary attempts to steal the model within a restricted access framework. While existing attacks demonstrate near-perfect…

The notion that collaborative machine learning can ensure privacy by just withholding the raw data is widely acknowledged to be flawed. Over the past seven years, the literature has revealed several privacy attacks that enable adversaries…

Training machine learning models on data from multiple entities without direct data sharing can unlock applications otherwise hindered by business, legal, or ethical constraints. In this work, we design and implement new privacy-preserving…

Jailbreaking is an emerging adversarial attack that bypasses the safety alignment deployed in off-the-shelf large language models (LLMs). A considerable amount of research exists proposing more effective jailbreak attacks, including the…

Cryptographic watermarking is a leading defense for attributing text generated by large language models (LLMs). Existing schemes, including KGW, Unigram, and DipMark, derive their security guarantees from the assumption that the underlying…