Adversarial Robustness in Machine Learning

Últimos artículos

• Fair Text Classification via Transferable Representations

  Thibaud Leteno, Michael Perrot, Charlotte Laclau, Antoine Gourru, Christophe Gravier · 8 de diciembre de 2025

  Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g., women and men) remains an open challenge. We propose an approach that extends the use of the Wasserstein Dependency Measure for learning unbiased neural text classifiers. …

• Beyond Data Filtering: Knowledge Localization for Capability Removal in LLMs

  Igor Shilov, Alex Cloud, Aryo Pradipta Gema, Jacob Goldman-Wetzler, Nina Panickssery, Henry Sleight, Erik Jones, Cem Anil · 8 de diciembre de 2025

  Large Language Models increasingly possess capabilities that carry dual-use risks. While data filtering has emerged as a pretraining-time mitigation, it faces significant challenges: labeling whether data is harmful is expensive at scale, and given improving sample efficiency with larger models, eve…

• Impugan: Learning Conditional Generative Models for Robust Data Imputation

  Zalish Mahmud, Anantaa Kotal, Aritran Piplai · 8 de diciembre de 2025

  Incomplete data are common in real-world applications. Sensors fail, records are inconsistent, and datasets collected from different sources often differ in scale, sampling rate, and quality. These differences create missing values that make it difficult to combine data and build reliable models. St…

• Edge-Only Universal Adversarial Attacks in Distributed Learning

  Giulio Rossolini, Tommaso Baldi, Alessandro Biondi, Giorgio Buttazzo · 8 de diciembre de 2025

  Distributed learning frameworks, which partition neural network models across multiple computing nodes, enhance efficiency in collaborative edge-cloud systems, but may also introduce new vulnerabilities to evasion attacks, often in the form of adversarial perturbations. In this work, we present a ne…

• ORFuzz: Fuzzing the "Other Side" of LLM Safety -- Testing Over-Refusal

  Haonan Zhang, Dongxia Wang, Yi Liu, Kexin Chen, Jiashui Wang, Xinlei Ying, Long Liu, Wenhai Wang · 8 de diciembre de 2025

  Large Language Models (LLMs) increasingly exhibit over-refusal - erroneously rejecting benign queries due to overly conservative safety measures - a critical functional flaw that undermines their reliability and usability. Current methods for testing this behavior are demonstrably inadequate, suffer…

• When unlearning is free: leveraging low influence points to reduce computational costs

  Anat Kleiman, Robert Fisher, Ben Deaner, Udi Wieder · 8 de diciembre de 2025

  As concerns around data privacy in machine learning grow, the ability to unlearn, or remove, specific data points from trained models becomes increasingly important. While state of the art unlearning methods have emerged in response, they typically treat all points in the forget set equally. In this…

• On the Bayes Inconsistency of Disagreement Discrepancy Surrogates

  Neil G. Marchant, Andrew C. Cullen, Feng Liu, Sarah M. Erfani · 8 de diciembre de 2025

  Deep neural networks often fail when deployed in real-world contexts due to distribution shift, a critical barrier to building safe and reliable systems. An emerging approach to address this problem relies on \emph{disagreement discrepancy} -- a measure of how the disagreement between two models cha…

• Taxonomy-Adaptive Moderation Model with Robust Guardrails for Large Language Models

  Mahesh Kumar Nandwana, Youngwan Lim, Joseph Liu, Alex Yang, Varun Notibala, Nishchaie Khanna · 8 de diciembre de 2025

  Large Language Models (LLMs) are typically aligned for safety during the post-training phase; however, they may still generate inappropriate outputs that could potentially pose risks to users. This challenge underscores the need for robust safeguards that operate across both model inputs and outputs…

• Spilling the Beans: Teaching LLMs to Self-Report Their Hidden Objectives

  Chloe Li, Mary Phuong, Daniel Tan · 8 de diciembre de 2025

  As AI systems become more capable of complex agentic tasks, they also become more capable of pursuing undesirable objectives and causing harm. Previous work has attempted to catch these unsafe instances by interrogating models directly about their objectives and behaviors. However, the main weakness…

• Credal and Interval Deep Evidential Classifications

  Michele Caprio, Shireen K. Manchingal, Fabio Cuzzolin · 8 de diciembre de 2025

  Uncertainty Quantification (UQ) presents a pivotal challenge in the field of Artificial Intelligence (AI), profoundly impacting decision-making, risk assessment and model reliability. In this paper, we introduce Credal and Interval Deep Evidential Classifications (CDEC and IDEC, respectively) as nov…

• LDLT $\mathcal{L}$-Lipschitz Network: Generalized Deep End-To-End Lipschitz Network Construction

  Marius F. R. Juston, Ramavarapu S. Sreenivas, Dustin Nottage, Ahmet Soylemezoglu · 8 de diciembre de 2025

  Deep residual networks (ResNets) have demonstrated outstanding success in computer vision tasks, attributed to their ability to maintain gradient flow through deep architectures. Simultaneously, controlling the Lipschitz constant in neural networks has emerged as an essential area of research to enh…

• Modular Jets for Supervised Pipelines: Diagnosing Mirage vs Identifiability

  Suman Sanyal · 8 de diciembre de 2025

  Classical supervised learning evaluates models primarily via predictive risk on hold-out data. Such evaluations quantify how well a function behaves on a distribution, but they do not address whether the internal decomposition of a model is uniquely determined by the data and evaluation design. In t…

• Matching Ranks Over Probability Yields Truly Deep Safety Alignment

  Jason Vega, Gagandeep Singh · 8 de diciembre de 2025

  A frustratingly easy technique known as the prefilling attack has been shown to effectively circumvent the safety alignment of frontier LLMs by simply prefilling the assistant response with an affirmative prefix before decoding. In response, recent work proposed a supervised fine-tuning (SFT) defens…

• Uncertainty-Aware Data-Efficient AI: An Information-Theoretic Perspective

  Osvaldo Simeone, Yaniv Romano · 8 de diciembre de 2025

  In context-specific applications such as robotics, telecommunications, and healthcare, artificial intelligence systems often face the challenge of limited training data. This scarcity introduces epistemic uncertainty, i.e., reducible uncertainty stemming from incomplete knowledge of the underlying d…

• Rethinking Sparse Autoencoders: Select-and-Project for Fairness and Control from Encoder Features Alone

  Antonio B\u{a}rb\u{a}lau, Cristian Daniel P\u{a}duraru, Teodor Poncu, Alexandru Tifrea, Elena Burceanu · 8 de diciembre de 2025

  Sparse Autoencoders (SAEs) are widely employed for mechanistic interpretability and model steering. Within this context, steering is by design performed by means of decoding altered SAE intermediate representations. This procedure essentially rewrites the original activations as a weighted sum of de…

• Robust Tabular Foundation Models

  Matthew Peroni, Franck Le, Vadim Sheinin · 4 de diciembre de 2025

  The development of tabular foundation models (TFMs) has accelerated in recent years, showing strong potential to outperform traditional ML methods for structured data. A key finding is that TFMs can be pretrained entirely on synthetic datasets, opening opportunities to design data generators that en…

• HalluGen: Synthesizing Realistic and Controllable Hallucinations for Evaluating Image Restoration

  Seunghoi Kim, Henry F. J. Tregidgo, Chen Jin, Matteo Figini, Daniel C. Alexander · 4 de diciembre de 2025

  Generative models are prone to hallucinations: plausible but incorrect structures absent in the ground truth. This issue is problematic in image restoration for safety-critical domains such as medical imaging, industrial inspection, and remote sensing, where such errors undermine reliability and tru…

• Context-Aware Hierarchical Learning: A Two-Step Paradigm towards Safer LLMs

  Tengyun Ma, Jiaqi Yao, Daojing He, Shihao Peng, Yu Li, Shaohui Liu, Zhuotao Tian · 4 de diciembre de 2025

  Large Language Models (LLMs) have emerged as powerful tools for diverse applications. However, their uniform token processing paradigm introduces critical vulnerabilities in instruction handling, particularly when exposed to adversarial scenarios. In this work, we identify and propose a novel class …

• Out-of-the-box: Black-box Causal Attacks on Object Detectors

  Melane Navaratnarajah, David A. Kelly, Hana Chockler · 4 de diciembre de 2025

  Adversarial perturbations are a useful way to expose vulnerabilities in object detectors. Existing perturbation methods are frequently white-box and architecture specific. More importantly, while they are often successful, it is rarely clear why they work. Insights into the mechanism of this success…

• In-Context Representation Hijacking

  Itay Yona, Amir Sarid, Michael Karasik, Yossi Gandelsman · 4 de diciembre de 2025

  We introduce \textbf{Doublespeak}, a simple \emph{in-context representation hijacking} attack against large language models (LLMs). The attack works by systematically replacing a harmful keyword (e.g., \textit{bomb}) with a benign token (e.g., \textit{carrot}) across multiple in-context examples, pr…

• DIQ-H: Evaluating Hallucination Persistence in VLMs Under Temporal Visual Degradation

  Zexin Lin, Hawen Wan, Yebin Zhong, Xiaoqiang · 4 de diciembre de 2025

  Vision-Language Models (VLMs) deployed in safety-critical applications such as autonomous driving must handle continuous visual streams under imperfect conditions. However, existing benchmarks focus on static, high-quality images and ignore temporal degradation and error propagation, which are criti…

• SELF: A Robust Singular Value and Eigenvalue Approach for LLM Fingerprinting

  Hanxiu Zhang, Yue Zheng · 4 de diciembre de 2025

  The protection of Intellectual Property (IP) in Large Language Models (LLMs) represents a critical challenge in contemporary AI research. While fingerprinting techniques have emerged as a fundamental mechanism for detecting unauthorized model usage, existing methods -- whether behavior-based or stru…

• MarkTune: Improving the Quality-Detectability Trade-off in Open-Weight LLM Watermarking

  Yizhou Zhao, Zhiwei Steven Wu, Adam Block · 4 de diciembre de 2025

  Watermarking aims to embed hidden signals in generated text that can be reliably detected when given access to a secret key. Open-weight language models pose acute challenges for such watermarking schemes because the inference-time interventions that dominate contemporary approaches cannot be enforc…

• Watermarks for Embeddings-as-a-Service Large Language Models

  Anudeex Shetty · 4 de diciembre de 2025

  Large Language Models (LLMs) have demonstrated exceptional capabilities in natural language understanding and generation. Based on these LLMs, businesses have started to provide Embeddings-as-a-Service (EaaS), offering feature extraction capabilities (in the form of text embeddings) that benefit dow…

• Many-to-One Adversarial Consensus: Exposing Multi-Agent Collusion Risks in AI-Based Healthcare

  Adeela Bashir, The Anh han, Zia Ush Shamszaman · 4 de diciembre de 2025

  The integration of large language models (LLMs) into healthcare IoT systems promises faster decisions and improved medical support. LLMs are also deployed as multi-agent teams to assist AI doctors by debating, voting, or advising on decisions. However, when multiple assistant agents interact, coordi…