485 citations · 1.5k across the 72 of their papers we have counts for
46 papers · 1 filter
LLMs Learn to Evade Latent Monitors from Prior Feedback Alone
Hugo Lyons Keenan, Christopher Leckie, Sarah Erfani
Latent space monitors aim to detect undesired behaviors in LLM agents by inspecting an agent's internal activations rather than its outputs. However, interactive monitoring creates…
VEX-Bench: Benchmarking Verification Complexity of LLM-Generated Misinformation
Hanxun Huang, Yutao Wu, Qizhou Wang +9
Large language models (LLMs) have made misinformation inexpensive to produce but not to verify, creating a growing asymmetry in the information ecosystem. Under tight time, labor,…
LAYERSCOPE: A Layerwise Characterization of Video and Multimodal Learned Representations
Sandra Arcos-Holzinger, Debashish Chakraborty, Rohita Mocharla +7
We propose LAYERSCOPE, a label-free, layerwise framework that aims to characterize a model's learned representations in video and multimodal settings. Evaluating downstream perform…
Semantic Robustness Certification for Vision-Language Models
Peiyu Yang, Paul Montague, Feng Liu +4
Vision-language models (VLMs) are now widely used in downstream tasks. However, real-world applications often expose VLMs to distribution shifts induced by semantic variation (e.g.…
TRACER: Persistent Regularization for Robust Multimodal Finetuning
Hesam Asadollahzadeh, Feng Liu, Christopher Leckie +1
Mainstream strategies for finetuning pretrained multimodal models often degrade out-of-distribution (OOD) robustness, a phenomenon known as catastrophic forgetting. In this paper,…
Mechanistic Anomaly Detection via Functional Attribution
Hugo Lyons Keenan, Christopher Leckie, Sarah Erfani
We can often verify the correctness of neural network outputs using ground truth labels, but we cannot reliably determine whether the output was produced by normal or anomalous int…
AudioMosaic: Contrastive Masked Audio Representation Learning
Hanxun Huang, Qizhou Wang, Xingjun Ma +3
Audio self-supervised learning (SSL) aims to learn general-purpose representations from large-scale unlabeled audio data. While recent advances have been driven mainly by generativ…
Fortifying Time Series: DTW-Certified Robust Anomaly Detection
Shijie Liu, Tansu Alpcan, Christopher Leckie +1
Time-series anomaly detection is critical for ensuring safety in high-stakes applications, where robustness is a fundamental requirement rather than a mere performance metric. Addr…
Toward Universal and Transferable Jailbreak Attacks on Vision-Language Models
Kaiyuan Cui, Yige Li, Yutao Wu +4
Vision-language models (VLMs) extend large language models (LLMs) with vision encoders, enabling text generation conditioned on both images and text. However, this multimodal integ…
Multi-level Certified Defense Against Poisoning Attacks in Offline Reinforcement Learning
Shijie Liu, Andrew C. Cullen, Paul Montague +2
Similar to other machine learning frameworks, Offline Reinforcement Learning (RL) is shown to be vulnerable to poisoning attacks, due to its reliance on externally sourced datasets…
Fox in the Henhouse: Supply-Chain Backdoor Attacks Against Reinforcement Learning
Shijie Liu, Andrew C. Cullen, Paul Montague +2
The current state-of-the-art backdoor attacks against Reinforcement Learning (RL) rely upon unrealistically permissive access models, that assume the attacker can read (or even wri…
HALO: Robust Out-of-Distribution Detection via Joint Optimisation
Hugo Lyons Keenan, Sarah Erfani, Christopher Leckie
Effective out-of-distribution (OOD) detection is crucial for the safe deployment of machine learning models in real-world scenarios. However, recent work has shown that OOD detecti…
Detecting Backdoor Samples in Contrastive Language Image Pretraining
Hanxun Huang, Sarah Erfani, Yige Li +2
Contrastive language-image pretraining (CLIP) has been found to be vulnerable to poisoning backdoor attacks where the adversary can achieve an almost perfect attack success rate on…
OIL-AD: An Anomaly Detection Framework for Sequential Decision Sequences
Chen Wang, Sarah Erfani, Tansu Alpcan +1
Anomaly detection in decision-making sequences is a challenging problem due to the complexity of normality representation learning and the sequential nature of the task. Most exist…
Unlearnable Examples For Time Series
Yujing Jiang, Xingjun Ma, Sarah Monazam Erfani +1
Unlearnable examples (UEs) refer to training samples modified to be unlearnable to Deep Neural Networks (DNNs). These examples are usually generated by adding error-minimizing nois…
Be Persistent: Towards a Unified Solution for Mitigating Shortcuts in Deep Learning
Hadi M. Dolatabadi, Sarah M. Erfani, Christopher Leckie
Deep neural networks (DNNs) are vulnerable to shortcut learning: rather than learning the intended task, they tend to draw inconclusive relationships between their inputs and outpu…
End-to-End Anti-Backdoor Learning on Images and Time Series
Yujing Jiang, Xingjun Ma, Sarah Monazam Erfani +2
Backdoor attacks present a substantial security concern for deep learning models, especially those utilized in applications critical to safety and security. These attacks manipulat…
LDReg: Local Dimensionality Regularized Self-Supervised Learning
Hanxun Huang, Ricardo J. G. B. Campello, Sarah Monazam Erfani +3
Representations learned via self-supervised learning (SSL) can be susceptible to dimensional collapse, where the learned representation subspace is of extremely low dimensionality…
It's Simplex! Disaggregating Measures to Improve Certified Robustness
Andrew C. Cullen, Paul Montague, Shijie Liu +2
Certified robustness circumvents the fragility of defences against adversarial attacks, by endowing model predictions with guarantees of class invariance for attacks up to a calcul…
Enhancing the Antidote: Improved Pointwise Certifications against Poisoning Attacks
Shijie Liu, Andrew C. Cullen, Paul Montague +2
Poisoning attacks can disproportionately influence model behaviour by making small changes to the training corpus. While defences against specific poisoning attacks do exist, they…
The Devil's Advocate: Shattering the Illusion of Unexploitable Data using Diffusion Models
Hadi M. Dolatabadi, Sarah Erfani, Christopher Leckie
Protecting personal data against exploitation of machine learning models is crucial. Recently, availability attacks have shown great promise to provide an extra layer of protection…
Et Tu Certifications: Robustness Certificates Yield Better Adversarial Examples
Andrew C. Cullen, Shijie Liu, Paul Montague +2
In guaranteeing the absence of adversarial examples in an instance's neighbourhood, certification mechanisms play an important role in demonstrating neural net robustness. In this…
Distilling Cognitive Backdoor Patterns within an Image
Hanxun Huang, Xingjun Ma, Sarah Erfani +1
This paper proposes a simple method to distill and detect backdoor patterns within an image: \emph{Cognitive Distillation} (CD). The idea is to extract the "minimal essence" from a…
Backdoor Attacks on Time Series: A Generative Approach
Yujing Jiang, Xingjun Ma, Sarah Monazam Erfani +1
Backdoor attacks have emerged as one of the major security threats to deep learning models as they can easily control the model's test-time predictions by pre-injecting a backdoor…