Related papers: Methodologies for Quantifying (Re-)randomization S…

Rust is a promising programming language that focuses on concurrency, usability, and security. It is used in production code by major industry players and got recommended by government bodies. Rust provides strong security guarantees…

Redundancy-based automated program repair (APR), which generates patches by referencing existing source code, has gained much attention since they are effective in repairing real-world bugs with good interpretability. However, since…

Hardware faults, specifically bit-flips in quantized weights, pose a severe reliability threat to Large Language Models (LLMs), often triggering catastrophic model collapses. We demonstrate that this vulnerability fundamentally stems from…

Although large language models (LLMs) are increasingly used in security-critical workflows, practitioners lack quantitative guidance on which safeguards are worth deploying. This paper introduces a decision-oriented framework and…

Large pre-trained Vision-Language Models (VLMs) like CLIP, despite having remarkable generalization ability, are highly vulnerable to adversarial examples. This work studies the adversarial robustness of VLMs from the novel perspective of…

Consumer Internet of Things (IoT) devices often leverage the local network to communicate with the corresponding companion app or other devices. This has benefits in terms of efficiency since it offloads the cloud. ENISA and NIST security…

Randomized benchmarking (RB) protocols are the most widely used methods for assessing the performance of quantum gates. However, the existing RB methods either do not scale to many qubits or cannot benchmark a universal gate set. Here, we…

Evaluating the programming robustness of large language models (LLMs) is paramount for ensuring their reliability in AI-based software development. However, adversarial attacks exhibit fundamental limitations that compromise fair robustness…

Since its inception, Rowhammer exploits have rapidly evolved into increasingly sophisticated threats compromising data integrity and the control flow integrity of victim processes. Nevertheless, it remains a challenge for an attacker to…

Typical quantum gate tomography protocols struggle with a self-consistency problem: the gate operation cannot be reconstructed without knowledge of the initial state and final measurement, but such knowledge cannot be obtained without…

Sequential recommendation aims to recommend the next item of users' interest based on their historical interactions. Recently, the self-attention mechanism has been adapted for sequential recommendation, and demonstrated state-of-the-art…

Robust training methods typically defend against specific attack types, such as Lp attacks with fixed budgets, and rarely account for the fact that defenders may encounter new attacks over time. A natural solution is to adapt the defended…

Regression testing assures software correctness after changes but is resource-intensive. Test Case Prioritisation (TCP) mitigates this by ordering tests to maximise early fault detection. Diversity-based TCP prioritises dissimilar tests,…

Randomized smoothing is a powerful tool for certifying robustness to adversarial perturbations, including poisoning attacks via randomized training and evasion attacks via randomized inference. Extending these guarantees to backdoor…

Vision-language models (VLMs), such as CLIP, have gained significant popularity as foundation models, with numerous fine-tuning methods developed to enhance performance on downstream tasks. However, due to their inherent vulnerability and…

Diversity coding is a network restoration technique which offers near-hitless restoration, while other state-of-the art techniques are significantly slower. Furthermore, the extra spare capacity requirement of diversity coding is…

Just in time defect prediction (JIT DP) leverages ML to identify defect-prone code commits, enabling quality assurance (QA) teams to allocate resources more efficiently by focusing on commits that are most likely to contain defects.…

Software redesign preserves functionality while improving quality attributes, but manual reuse of code and tests is costly and error-prone, especially in crossrepository redesigns. Focusing on static analyzers where cross-repo redesign…

Large language models (LLMs) rely on deterministic pseudorandom number generators (PRNGs) for autoregressive sampling, creating a critical supply-chain attack surface overlooked by existing defenses. We present SeedHijack, a backdoor attack…

This dissertation investigates how reinforcement learning (RL) methods can be designed to be safe, sample-efficient, and robust. Framed through the unifying perspective of contextual-bandit RL, the work addresses two major application…