When Rule Violations Are Rare: Chimera Training for Logical Anomaly Detection

• URL: http://arxiv.org/abs/2605.26171v1
• Date: Mon, 25 May 2026 02:52:36 GMT
• Title: When Rule Violations Are Rare: Chimera Training for Logical Anomaly Detection
• Abstract summary: We study anomaly detection in a setting where constraints are given as logical rules over learned visual concepts.<n>We propose a neural rule evaluator that compiles each constraint into a directed acyclic graph and learns feature-aware subtree gates.<n>Across CLEVRER, OpenImages, and VidOR, the resulting evaluator improves rule-level anomaly AUROC over independent-events and same-image semantic-training baselines.
• Score: 56.362776482614976
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Many practical anomalies are not merely rare inputs, but violations of semantic constraints: objects co-occur in structured ways, actions imply preconditions, and events satisfy temporal or relational regularities. We study anomaly detection in this setting, where constraints are given as logical rules over learned visual concepts, but real rule violations are rare or absent during training. We propose a neural rule evaluator that compiles each constraint into a directed acyclic graph and learns feature-aware subtree MLP gates for its internal logical operators. Each gate maps child features and edge-level negations to a parent representation and a rule-satisfaction probability, with intermediate supervision obtained from exact Boolean propagation over ground-truth concept labels. The key difficulty is that same-image training data often provide insufficient coverage of informative truth configurations and also allow shortcut solutions. To address this, we introduce chimera training: an operand-level counterfactual construction at the feature level. Instead of mixing input images, we concatenate subtree features from different samples; each operand keeps the hard truth label of the sample it came from, and the chimera target is obtained by applying the node's logical operator to those inherited labels. This supplies supervised logical counterexamples without requiring real anomalous images. Across CLEVRER, OpenImages, and VidOR, the resulting evaluator improves rule-level anomaly AUROC over independent-events and same-image semantic-training baselines, especially for compositional and relational rules. The method yields both scalar anomaly scores and rule-level attributions.

Related papers

• Training-Free Logical and Structural Anomaly Detection via Calibrated Fusion [4.800682857279701]
  We show that counting ability can be introduced into training-free anomaly detection without additional training or part-level supervision.<n>Our method achieves image-level AUROCs of 89.0 and 95.9 on logical and structural anomalies, respectively, yielding a 92.5 average.
  arXiv  Detail & Related papers  (2026-09-04T12:45:50Z)
• Differentiable Interval Bottlenecks for Interpretable Anomaly Detection in Numerical Data [0.5213778368155993]
  DIFFINT is an autoencoder whose latent bottleneck is structured as a set of soft, axis-aligned interval memberships.<n>It is the only interpretable detector in the statistically-tied leading cluster of seven methods.
  arXiv  Detail & Related papers  (2026-09-03T14:05:27Z)
• RAD: Rule-Augmented Relational Anomaly Detection [11.251069238986483]
  Anomaly detection is often applied to data stored in relational databases.<n>Most existing methods require flattening multiple tables into a single feature matrix.<n>We propose RAD, a rule-augmented relational anomaly detector.
  arXiv  Detail & Related papers  (2026-08-24T16:36:17Z)
• LC-ERD: Mining Latent Logic for Self-Evolving Reasoning via Consistency-Regulated Reward Decomposition [55.572260012037084]
  We introduce LC-ERD (Logic-Consistent Endogenous Reward Decomposition), a framework framing self-alignment as latent structure mining.<n>We derive a Variational Logic Potential by aggregating consensus from the model's Latent Logic Expertise.<n>Experiments show LC-ERD delivers a robust self-evolution path, uncovering trade-offs between logic consistency and accuracy.
  arXiv  Detail & Related papers  (2026-05-19T07:27:50Z)
• A Foundation Model for Zero-Shot Logical Rule Induction [0.0]
  We introduce Neural Rule Inducer (NRI), a pretrained model for zero-shot rule induction.<n>Rather than encoding literal identities, NRI represents literals using domain-agnostic statistical properties.<n>Product T-norm relaxation makes rule execution differentiable, allowing end-to-end training on prediction accuracy alone.
  arXiv  Detail & Related papers  (2026-05-06T13:45:31Z)
• ANDRE: An Attention-based Neuro-symbolic Differentiable Rule Extractor [6.477319771984887]
  Inductive Logic Programming aims to learn interpretable first-order rules from data.<n>Existing symbolic neuro-symbolic approaches struggle to scale to noisy and probabilistic settings.<n>This paper proposes an Attention-based Neuro-symbolic Differentiable Rule Extractor (ANDRE)<n>ANDRE replaces both rule templates and logical operators with fully differentiable, attention-driven conjunction and disjunction operators.
  arXiv  Detail & Related papers  (2026-05-05T18:35:06Z)
• Out of Context: Reliability in Multimodal Anomaly Detection Requires Contextual Inference [54.774686416991386]
  We argue that multimodal anomaly detection should be reframed as a cross-modal contextual inference problem.<n>This perspective has implications for model design, evaluation protocols, and benchmark construction.
  arXiv  Detail & Related papers  (2026-04-14T19:32:55Z)
• CodeCircuit: Toward Inferring LLM-Generated Code Correctness via Attribution Graphs [13.488544043942495]
  We aim to investigate whether the model's neural dynamics encode internally decodable signals that are predictive of logical validity during code generation.<n>By decomposing complex residual flows, we aim to identify the structural signatures that distinguish sound reasoning from logical failure.<n>Analysis across Python, C++, and Java confirms that intrinsic correctness signals are robust across diverse syntaxes.
  arXiv  Detail & Related papers  (2026-02-06T03:49:15Z)
• Interpretable Logical Anomaly Classification via Constraint Decomposition and Instruction Fine-Tuning [0.17722218114340835]
  We introduce Logical Anomaly Classification (LAC), a task that unifies anomaly detection and fine-grained violation classification in a single inference step.<n>To tackle LAC, we propose LogiCls, a vision-language framework that decomposes complex logical constraints into a sequence of verifiable subqueries.
  arXiv  Detail & Related papers  (2026-02-03T13:48:09Z)
• GeneralAD: Anomaly Detection Across Domains by Attending to Distorted Features [68.14842693208465]
  GeneralAD is an anomaly detection framework designed to operate in semantic, near-distribution, and industrial settings. We propose a novel self-supervised anomaly generation module that employs straightforward operations like noise addition and shuffling to patch features. We extensively evaluated our approach on ten datasets, achieving state-of-the-art results in six and on-par performance in the remaining.
  arXiv  Detail & Related papers  (2024-07-17T09:27:41Z)
• 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)
• Structured Prediction with Partial Labelling through the Infimum Loss [85.4940853372503]
  The goal of weak supervision is to enable models to learn using only forms of labelling which are cheaper to collect. This is a type of incomplete annotation where, for each datapoint, supervision is cast as a set of labels containing the real one. This paper provides a unified framework based on structured prediction and on the concept of infimum loss to deal with partial labelling.
  arXiv  Detail & Related papers  (2020-03-02T13:59:41Z)

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.