37 citations · 224 across the 83 of their papers we have counts for
55 papers · 1 filter
Quiet Triggers, Loud Footprints: A Tri-Space Measurement Study of Backdoor Stealthiness
Behrad Tajalli, Henk Berendsen, Xiaoyun Xu +1
Backdoor attacks on deep neural networks are commonly evaluated by attack success rate and benign accuracy, yet their stealthiness remains inconsistently defined and measured. Exis…
Beyond Scalar Probes: Exploiting Vector-Valued Outputs in ReLU Networks For Signature Extraction
Gorka Abad, Claude Carlet, Ermes Franch +2
We revisit cryptanalytic extraction of ReLU networks from a geometric and algebraic perspective. Rather than restricting attention to a single output component, we study the full v…
When the World Lies: Backdoor Attacks on Latent World Models for Downstream Control
Roberto Riaño, Gorka Abad, Stjepan Picek +1
Pretrained world models, learned simulators that encode an observation into a latent state and predict how it evolves under actions, are beginning to be reused as off-the-shelf dyn…
SIR: Self-improving Red-teaming for Compute Use Agents
Chen Xiong, Zhiyuan He, Pin-Yu Chen +2
Computer-use agents (CUAs) are agents powered by vision-language models (VLMs) that perceive a screen and operate an operating system through mouse, keyboard, and terminal interact…
WoE Wrote It? Watermarking Mixture-of-Experts LLMs for Black-Box Text Provenance
Jona te Lintelo, Lichao Wu, Stjepan Picek
Large Language Model (LLM) watermarks provide a mechanism for text provenance, enabling model owners to identify machine-generated content and attribute it to a specific watermarke…
MeMark: Membrane-Space Watermarking for Spiking Neural Networks
Roberto Riaño, Gorka Abad, Stjepan Picek +1
Spiking Neural Networks (SNNs) are increasingly distributed as pretrained checkpoints and reused as backbones for new tasks. However, current SNN watermarks are mainly verified aga…
From Verdict to Diagnosis: Attributable Security Review of Pull Requests
Zhuo Chen, Boyang Wang, Xiyue Zhang +4
Automated code reviewers are increasingly used as gates on pull requests (PRs), yet evaluations measure whether they block a malicious change. A block may be triggered by an unrela…
Temporal Poisoning: Clean-Label Backdoors via Event Redistribution in SNNs
Roberto Riaño, Gorka Abad, Stjepan Picek +1
Backdoor attacks on Spiking Neural Networks (SNNs) have primarily assumed dirty-label poisoning, in which triggered training samples are relabeled to an attacker-selected class. We…
Activating Latent Security Knowledge through LLM-Guided Risk Analysis for Secure Code Generation
Xiaoyun Xu, Lichao Wu, Jona te Lintelo +3
Large language models are pretrained on extensive software and security corpora, yet they frequently generate functionally correct code containing well-known vulnerabilities. Exist…
You Snooze, You Lose: Automatic Safety Alignment Restoration through Neural Weight Translation
Marco Arazzi, Vignesh Kumar Kembu, Antonino Nocera +2
Public repositories now distribute thousands of specialized Low-Rank Adaptation (LoRA) modules, but a third-party adapter routinely arrives without the refusal behavior the same ad…
MASCing: Configurable Mixture-of-Experts Behavior via Activation Steering Masks
Jona te Lintelo, Lichao Wu, Marina Krček +2
Mixture-of-Experts (MoE) architectures in Large Language Models (LLMs) have significantly reduced inference costs through sparse activation. However, this sparse activation paradig…
Backdoor Attacks on Decentralised Post-Training
Oğuzhan Ersoy, Nikolay Blagoev, Jona te Lintelo +3
Decentralised post-training of large language models utilises data and pipeline parallelism techniques to split the data and the model. Unfortunately, decentralised post-training c…
Large Language Lobotomy: Jailbreaking Mixture-of-Experts via Expert Silencing
Jona te Lintelo, Lichao Wu, Stjepan Picek
The rapid adoption of Mixture-of-Experts (MoE) architectures marks a major shift in the deployment of Large Language Models (LLMs). MoE LLMs improve scaling efficiency by activatin…
GoodVibe: Security-by-Vibe for LLM-Based Code Generation
Maximilian Thang, Lichao Wu, Sasha Behrouzi +4
Large language models (LLMs) are increasingly used for code generation in fast, informal development workflows, often referred to as vibe coding, where speed and convenience are pr…
IDEM Enough? Evolving Highly Nonlinear Idempotent Boolean Functions
Claude Carlet, Marko Ðurasevic, Domagoj Jakobovic +2
Idempotent Boolean functions form a highly structured subclass of Boolean functions that is closely related to rotation symmetry under a normal-basis representation and to invarian…
Emerging Threats and Countermeasures in Neuromorphic Systems: A Survey
Pablo Sorrentino, Stjepan Picek, Ihsen Alouani +5
Neuromorphic computing mimics brain-inspired mechanisms through spiking neurons and energy-efficient processing, offering a pathway to efficient in-memory computing (IMC). However,…
GateBreaker: Gate-Guided Attacks on Mixture-of-Expert LLMs
Lichao Wu, Sasha Behrouzi, Mohamadreza Rostami +2
Mixture-of-Experts (MoE) architectures have advanced the scaling of Large Language Models (LLMs) by activating only a sparse subset of parameters per input, enabling state-of-the-a…
SoK: The Last Line of Defense: On Backdoor Defense Evaluation
Gorka Abad, Marina Krček, Stefanos Koffas +7
Backdoor attacks pose a significant threat to deep learning models by implanting hidden vulnerabilities that can be activated by malicious inputs. While numerous defenses have been…
NoMod: A Non-modular Attack on Module Learning With Errors
Cristian Bassotto, Ermes Franch, Marina Krček +1
The advent of quantum computing threatens classical public-key cryptography, motivating NIST's adoption of post-quantum schemes such as those based on the Module Learning With Erro…
NeuroStrike: Neuron-Level Attacks on Aligned LLMs
Lichao Wu, Sasha Behrouzi, Mohamadreza Rostami +3
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…
BadPatches: Routing-Aware Backdoor Attacks on Vision Mixture-of-Experts
Jona te Lintelo, Cedric Chan, Stjepan Picek
Mixture-of-Experts (MoE) architectures have gained significant traction for reducing computational costs in deep neural networks by activating only a sparse subset of parameters du…
Hardware-Enabled Mechanisms for Verifying Responsible AI Development
Aidan O'Gara, Gabriel Kulp, Will Hodgkins +7
Advancements in AI capabilities, driven in large part by scaling up computing resources used for AI training, have created opportunities to address major global challenges but also…
Interpreting Emergent Features in Deep Learning-based Side-channel Analysis
Sengim Karayalçin, Marina Krček, Stjepan Picek
Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the s…
Towards Backdoor Stealthiness in Model Parameter Space
Xiaoyun Xu, Zhuoran Liu, Stefanos Koffas +1
Recent research on backdoor stealthiness focuses mainly on indistinguishable triggers in input space and inseparable backdoor representations in feature space, aiming to circumvent…