Searcharxiv⌕ Search

arXiv subjects

Tamas Abraham

Publications and source records attributed to Tamas Abraham.

At least 19 recordsLinked to original sources
arXiv2026

Adaptive Subspace Projection for Generative Personalization

Generative personalization often suffers from the semantic collapsing problem (SCP), where a learned personalized concept overpowers the rest of the text prompt, causing the model to ignore important contextual details. To address this, we first analyze the underlying cause, revealing that the semantic drift responsible for SCP is not random but is concentrated within a specific low-dimensional subspace. We also discover that the personalization process perturbs the embedding of the original base concept, making it an unstable reference point. Based on these insights, we introduce Test-time Embedding Adjustment with Adaptive Subspace Projection (AdaptSP), a training-free method that uses the stable, pre-trained embedding as an anchor. AdaptSP isolates the semantic drift and projects it onto the identified subspace, performing a precise adjustment that mitigates SCP while maintaining the subject identity. Our experiments show that this targeted approach significantly improves prompt fidelity and contextual alignment.

cs.CV↗
arXiv2026

MCBench: A Multicontext Safety Assessment Benchmark for Omni Large Language Models

Existing multimodal safety benchmarks focus solely on visual inputs and cannot assess Omni Large Language Models (LLMs) that process vision, audio, and text. We introduce MCBench, a benchmark with 1196 scenarios spanning four safety categories that require integrating multiple modalities for accurate safety assessment. Each unsafe scenario is paired with a minimally different safe counterpart to assess model sensitivity. Our evaluations of state-of-the-art models reveal significant challenges. Omni LLMs struggle with subtle or non-physical risks but perform better when salient visual or acoustic cues are present. Analysis of reasoning traces shows that, although models can extract modality-specific information, they often fail to integrate these cues effectively for safety judgments. Our findings reveal that current Omni LLMs lack robust cross-modal reasoning in safety-critical settings, underscoring the need for improved architectures and training strategies for multimodal safety.

cs.CL↗
arXiv2025

Certified but Fooled! Breaking Certified Defences with Ghost Certificates

Certified defenses promise provable robustness guarantees. We study the malicious exploitation of probabilistic certification frameworks to better understand the limits of guarantee provisions. Now, the objective is to not only mislead a classifier, but also manipulate the certification process to generate a robustness guarantee for an adversarial input certificate spoofing. A recent study in ICLR demonstrated that crafting large perturbations can shift inputs far into regions capable of generating a certificate for an incorrect class. Our study investigates if perturbations needed to cause a misclassification and yet coax a certified model into issuing a deceptive, large robustness radius for a target class can still be made small and imperceptible. We explore the idea of region-focused adversarial examples to craft imperceptible perturbations, spoof certificates and achieve certification radii larger than the source class ghost certificates. Extensive evaluations with the ImageNet demonstrate the ability to effectively bypass state-of-the-art certified defenses such as Densepure. Our work underscores the need to better understand the limits of robustness certification methods.

cs.LG↗
arXiv2025

The Social Cost of Intelligence: Emergence, Propagation, and Amplification of Stereotypical Bias in Multi-Agent Systems

Bias in large language models (LLMs) remains a persistent challenge, often leading to stereotyping and unfair treatment across social groups. While prior work has mainly focused on individual LLMs, the emergence of multi-agent systems (MAS), where multiple LLMs collaborate and communicate, introduces new and underexplored dynamics in how bias emerges, propagates, and amplifies. To systematically investigate these dynamics, we propose a simple evaluation framework with three agent-level metrics that quantify bias emergence, propagation, and amplification throughout multi-agent interaction. We evaluate MAS across three bias benchmarks under varying LLM backbones, social-group configurations, communication behaviors, and adversarial settings. Our results show that communication can trigger up to 70\% new bias emergence, propagate bias across over 80\% of agents, and amplify stereotypes by more than 3$\times$. We further find that denser and competitive communication generally increases bias. Finally, we demonstrate that MAS are highly vulnerable to simple bias injection attacks, and existing defense strategies provide only limited protection. Our findings provide important insights into the fairness and robustness of multi-agent LLM systems.

cs.MA↗
arXiv2025

Mitigating Semantic Collapse in Generative Personalization with Test-Time Embedding Adjustment

In this paper, we investigate the semantic collapsing problem in generative personalization, an under-explored topic where the learned visual concept ($V$) gradually shifts from its original textual meaning and comes to dominate other concepts in multi-concept input prompts. This issue not only reduces the semantic richness of complex input prompts like "a photo of $V$ wearing glasses and playing guitar" into simpler, less contextually rich forms such as "a photo of $V$" but also leads to simplified output images that fail to capture the intended concept. We identify the root cause as unconstrained optimisation, which allows the learned embedding $V$ to drift arbitrarily in the embedding space, both in direction and magnitude. To address this, we propose a simple yet effective training-free method that adjusts the magnitude and direction of pre-trained embedding at inference time, effectively mitigating the semantic collapsing problem. Our method is broadly applicable across different personalization methods and demonstrates significant improvements in text-image alignment in diverse use cases. Our code is anonymously published at https://github.com/tuananhbui89/Embedding-Adjustment

cs.LG↗
arXiv2025

Fantastic Targets for Concept Erasure in Diffusion Models and Where To Find Them

Concept erasure has emerged as a promising technique for mitigating the risk of harmful content generation in diffusion models by selectively unlearning undesirable concepts. The common principle of previous works to remove a specific concept is to map it to a fixed generic concept, such as a neutral concept or just an empty text prompt. In this paper, we demonstrate that this fixed-target strategy is suboptimal, as it fails to account for the impact of erasing one concept on the others. To address this limitation, we model the concept space as a graph and empirically analyze the effects of erasing one concept on the remaining concepts. Our analysis uncovers intriguing geometric properties of the concept space, where the influence of erasing a concept is confined to a local region. Building on this insight, we propose the Adaptive Guided Erasure (AGE) method, which \emph{dynamically} selects optimal target concepts tailored to each undesirable concept, minimizing unintended side effects. Experimental results show that AGE significantly outperforms state-of-the-art erasure methods on preserving unrelated concepts while maintaining effective erasure performance. Our code is published at {https://github.com/tuananhbui89/Adaptive-Guided-Erasure}.

cs.LG↗
arXiv2024

Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation

Diffusion models excel at generating visually striking content from text but can inadvertently produce undesirable or harmful content when trained on unfiltered internet data. A practical solution is to selectively removing target concepts from the model, but this may impact the remaining concepts. Prior approaches have tried to balance this by introducing a loss term to preserve neutral content or a regularization term to minimize changes in the model parameters, yet resolving this trade-off remains challenging. In this work, we propose to identify and preserving concepts most affected by parameter changes, termed as \textit{adversarial concepts}. This approach ensures stable erasure with minimal impact on the other concepts. We demonstrate the effectiveness of our method using the Stable Diffusion model, showing that it outperforms state-of-the-art erasure methods in eliminating unwanted content while maintaining the integrity of other unrelated elements. Our code is available at https://github.com/tuananhbui89/Erasing-Adversarial-Preservation.

cs.LG↗
arXiv2024

On the Credibility of Backdoor Attacks Against Object Detectors in the Physical World

Object detectors are vulnerable to backdoor attacks. In contrast to classifiers, detectors possess unique characteristics, architecturally and in task execution; often operating in challenging conditions, for instance, detecting traffic signs in autonomous cars. But, our knowledge dominates attacks against classifiers and tests in the "digital domain". To address this critical gap, we conducted an extensive empirical study targeting multiple detector architectures and two challenging detection tasks in real-world settings: traffic signs and vehicles. Using the diverse, methodically collected videos captured from driving cars and flying drones, incorporating physical object trigger deployments in authentic scenes, we investigated the viability of physical object-triggered backdoor attacks in application settings. Our findings revealed 8 key insights. Importantly, the prevalent "digital" data poisoning method for injecting backdoors into models does not lead to effective attacks against detectors in the real world, although proven effective in classification tasks. We construct a new, cost-efficient attack method, dubbed MORPHING, incorporating the unique nature of detection tasks; ours is remarkably successful in injecting physical object-triggered backdoors, even capable of poisoning triggers with clean label annotations or invisible triggers without diminishing the success of physical object triggered backdoors. We discovered that the defenses curated are ill-equipped to safeguard detectors against such attacks. To underscore the severity of the threat and foster further research, we, for the first time, release an extensive video test set of real-world backdoor attacks. Our study not only establishes the credibility and seriousness of this threat but also serves as a clarion call to the research community to advance backdoor defenses in the context of object detection.

cs.CR↗
arXiv2024

EaTVul: ChatGPT-based Evasion Attack Against Software Vulnerability Detection

Recently, deep learning has demonstrated promising results in enhancing the accuracy of vulnerability detection and identifying vulnerabilities in software. However, these techniques are still vulnerable to attacks. Adversarial examples can exploit vulnerabilities within deep neural networks, posing a significant threat to system security. This study showcases the susceptibility of deep learning models to adversarial attacks, which can achieve 100% attack success rate (refer to Table 5). The proposed method, EaTVul, encompasses six stages: identification of important samples using support vector machines, identification of important features using the attention mechanism, generation of adversarial data based on these features using ChatGPT, preparation of an adversarial attack pool, selection of seed data using a fuzzy genetic algorithm, and the execution of an evasion attack. Extensive experiments demonstrate the effectiveness of EaTVul, achieving an attack success rate of more than 83% when the snippet size is greater than 2. Furthermore, in most cases with a snippet size of 4, EaTVul achieves a 100% attack success rate. The findings of this research emphasize the necessity of robust defenses against adversarial attacks in software vulnerability detection.

cs.CR↗
arXiv2024

Bayesian Learned Models Can Detect Adversarial Malware For Free

The vulnerability of machine learning-based malware detectors to adversarial attacks has prompted the need for robust solutions. Adversarial training is an effective method but is computationally expensive to scale up to large datasets and comes at the cost of sacrificing model performance for robustness. We hypothesize that adversarial malware exploits the low-confidence regions of models and can be identified using epistemic uncertainty of ML approaches -- epistemic uncertainty in a machine learning-based malware detector is a result of a lack of similar training samples in regions of the problem space. In particular, a Bayesian formulation can capture the model parameters' distribution and quantify epistemic uncertainty without sacrificing model performance. To verify our hypothesis, we consider Bayesian learning approaches with a mutual information-based formulation to quantify uncertainty and detect adversarial malware in Android, Windows domains and PDF malware. We found, quantifying uncertainty through Bayesian learning methods can defend against adversarial malware. In particular, Bayesian models: (1) are generally capable of identifying adversarial malware in both feature and problem space, (2) can detect concept drift by measuring uncertainty, and (3) with a diversity-promoting approach (or better posterior approximations) lead to parameter instances from the posterior to significantly enhance a detectors' ability.

cs.CR↗
arXiv2024

Hiding and Recovering Knowledge in Text-to-Image Diffusion Models via Learnable Prompts

Diffusion models have demonstrated remarkable capability in generating high-quality visual content from textual descriptions. However, since these models are trained on large-scale internet data, they inevitably learn undesirable concepts, such as sensitive content, copyrighted material, and harmful or unethical elements. While previous works focus on permanently removing such concepts, this approach is often impractical, as it can degrade model performance and lead to irreversible loss of information. In this work, we introduce a novel concept-hiding approach that makes unwanted concepts inaccessible to public users while allowing controlled recovery when needed. Instead of erasing knowledge from the model entirely, we incorporate a learnable prompt into the cross-attention module, acting as a secure memory that suppresses the generation of hidden concepts unless a secret key is provided. This enables flexible access control -- ensuring that undesirable content cannot be easily generated while preserving the option to reinstate it under restricted conditions. Our method introduces a new paradigm where concept suppression and controlled recovery coexist, which was not feasible in prior works. We validate its effectiveness on the Stable Diffusion model, demonstrating that hiding concepts mitigate the risks of permanent removal while maintaining the model's overall capability.

cs.LG↗
arXiv2023

Feature-Space Bayesian Adversarial Learning Improved Malware Detector Robustness

We present a new algorithm to train a robust malware detector. Modern malware detectors rely on machine learning algorithms. Now, the adversarial objective is to devise alterations to the malware code to decrease the chance of being detected whilst preserving the functionality and realism of the malware. Adversarial learning is effective in improving robustness but generating functional and realistic adversarial malware samples is non-trivial. Because: i) in contrast to tasks capable of using gradient-based feedback, adversarial learning in a domain without a differentiable mapping function from the problem space (malware code inputs) to the feature space is hard; and ii) it is difficult to ensure the adversarial malware is realistic and functional. This presents a challenge for developing scalable adversarial machine learning algorithms for large datasets at a production or commercial scale to realize robust malware detectors. We propose an alternative; perform adversarial learning in the feature space in contrast to the problem space. We prove the projection of perturbed, yet valid malware, in the problem space into feature space will always be a subset of adversarials generated in the feature space. Hence, by generating a robust network against feature-space adversarial examples, we inherently achieve robustness against problem-space adversarial examples. We formulate a Bayesian adversarial learning objective that captures the distribution of models for improved robustness. We prove that our learning method bounds the difference between the adversarial risk and empirical risk explaining the improved robustness. We show that adversarially trained BNNs achieve state-of-the-art robustness. Notably, adversarially trained BNNs are robust against stronger attacks with larger attack budgets by a margin of up to 15% on a recent production-scale malware dataset of more than 20 million samples.

cs.CR↗
arXiv2022

Transferable Graph Backdoor Attack

Graph Neural Networks (GNNs) have achieved tremendous success in many graph mining tasks benefitting from the message passing strategy that fuses the local structure and node features for better graph representation learning. Despite the success of GNNs, and similar to other types of deep neural networks, GNNs are found to be vulnerable to unnoticeable perturbations on both graph structure and node features. Many adversarial attacks have been proposed to disclose the fragility of GNNs under different perturbation strategies to create adversarial examples. However, vulnerability of GNNs to successful backdoor attacks was only shown recently. In this paper, we disclose the TRAP attack, a Transferable GRAPh backdoor attack. The core attack principle is to poison the training dataset with perturbation-based triggers that can lead to an effective and transferable backdoor attack. The perturbation trigger for a graph is generated by performing the perturbation actions on the graph structure via a gradient based score matrix from a surrogate model. Compared with prior works, TRAP attack is different in several ways: i) it exploits a surrogate Graph Convolutional Network (GCN) model to generate perturbation triggers for a blackbox based backdoor attack; ii) it generates sample-specific perturbation triggers which do not have a fixed pattern; and iii) the attack transfers, for the first time in the context of GNNs, to different GNN models when trained with the forged poisoned training dataset. Through extensive evaluations on four real-world datasets, we demonstrate the effectiveness of the TRAP attack to build transferable backdoors in four different popular GNNs using four real-world datasets.

cs.CR↗
arXiv2022

Learning to Attack with Fewer Pixels: A Probabilistic Post-hoc Framework for Refining Arbitrary Dense Adversarial Attacks

Deep neural network image classifiers are reported to be susceptible to adversarial evasion attacks, which use carefully crafted images created to mislead a classifier. Many adversarial attacks belong to the category of dense attacks, which generate adversarial examples by perturbing all the pixels of a natural image. To generate sparse perturbations, sparse attacks have been recently developed, which are usually independent attacks derived by modifying a dense attack's algorithm with sparsity regularisations, resulting in reduced attack efficiency. In this paper, we aim to tackle this task from a different perspective. We select the most effective perturbations from the ones generated from a dense attack, based on the fact we find that a considerable amount of the perturbations on an image generated by dense attacks may contribute little to attacking a classifier. Accordingly, we propose a probabilistic post-hoc framework that refines given dense attacks by significantly reducing the number of perturbed pixels but keeping their attack power, trained with mutual information maximisation. Given an arbitrary dense attack, the proposed model enjoys appealing compatibility for making its adversarial images more realistic and less detectable with fewer perturbations. Moreover, our framework performs adversarial attacks much faster than existing sparse attacks.

cs.CV↗
arXiv2022

Improving Ensemble Robustness by Collaboratively Promoting and Demoting Adversarial Robustness

Ensemble-based adversarial training is a principled approach to achieve robustness against adversarial attacks. An important technique of this approach is to control the transferability of adversarial examples among ensemble members. We propose in this work a simple yet effective strategy to collaborate among committee models of an ensemble model. This is achieved via the secure and insecure sets defined for each model member on a given sample, hence help us to quantify and regularize the transferability. Consequently, our proposed framework provides the flexibility to reduce the adversarial transferability as well as to promote the diversity of ensemble members, which are two crucial factors for better robustness in our ensemble approach. We conduct extensive and comprehensive experiments to demonstrate that our proposed method outperforms the state-of-the-art ensemble baselines, at the same time can detect a wide range of adversarial examples with a nearly perfect accuracy. Our code is available at: https://github.com/tuananhbui89/Crossing-Collaborative-Ensemble.

cs.CV↗
arXiv2020

Adversarial Reinforcement Learning under Partial Observability in Autonomous Computer Network Defence

Recent studies have demonstrated that reinforcement learning (RL) agents are susceptible to adversarial manipulation, similar to vulnerabilities previously demonstrated in the supervised learning setting. While most existing work studies the problem in the context of computer vision or console games, this paper focuses on reinforcement learning in autonomous cyber defence under partial observability. We demonstrate that under the black-box setting, where the attacker has no direct access to the target RL model, causative attacks---attacks that target the training process---can poison RL agents even if the attacker only has partial observability of the environment. In addition, we propose an inversion defence method that aims to apply the opposite perturbation to that which an attacker might use to generate their adversarial samples. Our experimental results illustrate that the countermeasure can effectively reduce the impact of the causative attack, while not significantly affecting the training process in non-attack scenarios.

stat.ML↗
arXiv2020

Improving Adversarial Robustness by Enforcing Local and Global Compactness

The fact that deep neural networks are susceptible to crafted perturbations severely impacts the use of deep learning in certain domains of application. Among many developed defense models against such attacks, adversarial training emerges as the most successful method that consistently resists a wide range of attacks. In this work, based on an observation from a previous study that the representations of a clean data example and its adversarial examples become more divergent in higher layers of a deep neural net, we propose the Adversary Divergence Reduction Network which enforces local/global compactness and the clustering assumption over an intermediate layer of a deep neural network. We conduct comprehensive experiments to understand the isolating behavior of each component (i.e., local/global compactness and the clustering assumption) and compare our proposed model with state-of-the-art adversarial training methods. The experimental results demonstrate that augmenting adversarial training with our proposed components can further improve the robustness of the network, leading to higher unperturbed and adversarial predictive performances.

cs.LG↗
arXiv2019

Perturbations are not Enough: Generating Adversarial Examples with Spatial Distortions

Deep neural network image classifiers are reported to be susceptible to adversarial evasion attacks, which use carefully crafted images created to mislead a classifier. Recently, various kinds of adversarial attack methods have been proposed, most of which focus on adding small perturbations to input images. Despite the success of existing approaches, the way to generate realistic adversarial images with small perturbations remains a challenging problem. In this paper, we aim to address this problem by proposing a novel adversarial method, which generates adversarial examples by imposing not only perturbations but also spatial distortions on input images, including scaling, rotation, shear, and translation. As humans are less susceptible to small spatial distortions, the proposed approach can produce visually more realistic attacks with smaller perturbations, able to deceive classifiers without affecting human predictions. We learn our method by amortized techniques with neural networks and generate adversarial examples efficiently by a forward pass of the networks. Extensive experiments on attacking different types of non-robustified classifiers and robust classifiers with defence show that our method has state-of-the-art performance in comparison with advanced attack parallels.

cs.LG↗