Mechanistic Anomaly Detection via Functional Attribution

• URL: http://arxiv.org/abs/2604.18970v1
• Date: Tue, 21 Apr 2026 01:39:57 GMT
• Title: Mechanistic Anomaly Detection via Functional Attribution
• Abstract summary: We implement a functional attribution problem for anomaly detection in neural networks.<n>For backdoors in vision models, our method achieves state-of-the-art detection on BackdoorBench.<n>Our results establish functional attribution as an effective, modality-agnostic tool for detecting anomalous behavior in deployed models.
• Score: 6.1937472685875
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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 internal mechanisms. Mechanistic anomaly detection (MAD) aims to flag these cases, but existing methods either depend on latent space analysis, which is vulnerable to obfuscation, or are specific to particular architectures and modalities. We reframe MAD as a functional attribution problem: asking to what extent samples from a trusted set can explain the model's output, where attribution failure signals anomalous behavior. We operationalize this using influence functions, measuring functional coupling between test samples and a small reference set via parameter-space sampling. We evaluate across multiple anomaly types and modalities. For backdoors in vision models, our method achieves state-of-the-art detection on BackdoorBench, with an average Defense Effectiveness Rating (DER) of 0.93 across seven attacks and four datasets (next best 0.83). For LLMs, we similarly achieve a significant improvement over baselines for several backdoor types, including on explicitly obfuscated models. Beyond backdoors, our method can detect adversarial and out-of-distribution samples, and distinguishes multiple anomalous mechanisms within a single model. Our results establish functional attribution as an effective, modality-agnostic tool for detecting anomalous behavior in deployed models.

Related papers

• Rethinking Structural Anomaly Detection: From Decision Boundaries to Projection Operators [51.771100367755544]
  We introduce a geometric perspective to address the mismatch between the inductive bias of existing methods and the structure of the data.<n>Specifically, we learn a projection operator onto the manifold of normal samples and define a sample as anomalous if it is altered by this projection.<n> Notably, it provides a unifying interpretation of reconstruction-based methods by explaining their success and failure in terms of projection quality.
  arXiv  Detail & Related papers  (2026-06-13T12:37:20Z)
• AnomalyAgent: Training-Free Agentic Models for Zero-/Few-Shot Anomaly Detection [43.187781277950386]
  We propose textbfAnomalyAgent, a training-free agentic framework for anomaly detection.<n>Key ingredients include textbf1) a comprehensive anomaly-centric toolset, and textbf2) a customized memory module.<n>Experiment results demonstrate that our AnomalyAgent achieves substantially better performance compared to training-free VLM-based AD and generic agentic methods.
  arXiv  Detail & Related papers  (2026-05-28T16:05:42Z)
• Feature-Aware Test Generation for Deep Learning Models [0.5368630420272898]
  We introduce Detect, a feature-aware test generation framework for vision-based deep learning (DL) models.<n>It generates inputs by perturbing disentangled semantic attributes within the latent space.<n>It identifies which features lead to behavior shifts and uses a vision-language model for semantic attribution.
  arXiv  Detail & Related papers  (2026-01-20T15:41:06Z)
• Membership Inference Attack with Partial Features [14.336020797572361]
  We study an inference scenario where the adversary observes only partial features of each sample.<n>We propose MRAD (Memory-guided Reconstruction and Anomaly Detection), a two-stage attack framework.<n>MRAD is effective across a range of datasets, and maintains compatibility with various off-the-shelf anomaly detection techniques.
  arXiv  Detail & Related papers  (2025-08-08T11:56:13Z)
• CLIP Meets Diffusion: A Synergistic Approach to Anomaly Detection [49.11819337853632]
  Anomaly detection is a complex problem due to the ambiguity in defining anomalies, the diversity of anomaly types, and the scarcity of training data.<n>We propose CLIPfusion, a method that leverages both discriminative and generative foundation models.<n>We believe that our method underscores the effectiveness of multi-modal and multi-model fusion in tackling the multifaceted challenges of anomaly detection.
  arXiv  Detail & Related papers  (2025-06-13T13:30:15Z)
• Can I trust my anomaly detection system? A case study based on explainable AI [0.4416503115535552]
  This case study explores the robustness of an anomaly detection system based on variational autoencoder generative models. The goal is to get a different perspective on the real performances of anomaly detectors that use reconstruction differences.
  arXiv  Detail & Related papers  (2024-07-29T12:39:07Z)
• Lazy Layers to Make Fine-Tuned Diffusion Models More Traceable [70.77600345240867]
  A novel arbitrary-in-arbitrary-out (AIAO) strategy makes watermarks resilient to fine-tuning-based removal. Unlike the existing methods of designing a backdoor for the input/output space of diffusion models, in our method, we propose to embed the backdoor into the feature space of sampled subpaths. Our empirical studies on the MS-COCO, AFHQ, LSUN, CUB-200, and DreamBooth datasets confirm the robustness of AIAO.
  arXiv  Detail & Related papers  (2024-05-01T12:03:39Z)
• Adversarial Examples Detection with Enhanced Image Difference Features based on Local Histogram Equalization [20.132066800052712]
  We propose an adversarial example detection framework based on a high-frequency information enhancement strategy. This framework can effectively extract and amplify the feature differences between adversarial examples and normal examples.
  arXiv  Detail & Related papers  (2023-05-08T03:14:01Z)
• Self-Supervised Training with Autoencoders for Visual Anomaly Detection [61.62861063776813]
  We focus on a specific use case in anomaly detection where the distribution of normal samples is supported by a lower-dimensional manifold. We adapt a self-supervised learning regime that exploits discriminative information during training but focuses on the submanifold of normal examples. We achieve a new state-of-the-art result on the MVTec AD dataset -- a challenging benchmark for visual anomaly detection in the manufacturing domain.
  arXiv  Detail & Related papers  (2022-06-23T14:16:30Z)
• Shaping Deep Feature Space towards Gaussian Mixture for Visual Classification [74.48695037007306]
  We propose a Gaussian mixture (GM) loss function for deep neural networks for visual classification. With a classification margin and a likelihood regularization, the GM loss facilitates both high classification performance and accurate modeling of the feature distribution. The proposed model can be implemented easily and efficiently without using extra trainable parameters.
  arXiv  Detail & Related papers  (2020-11-18T03:32:27Z)
• Unsupervised Anomaly Detection with Adversarial Mirrored AutoEncoders [51.691585766702744]
  We propose a variant of Adversarial Autoencoder which uses a mirrored Wasserstein loss in the discriminator to enforce better semantic-level reconstruction. We put forward an alternative measure of anomaly score to replace the reconstruction-based metric. Our method outperforms the current state-of-the-art methods for anomaly detection on several OOD detection benchmarks.
  arXiv  Detail & Related papers  (2020-03-24T08:26:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.