Learning to Deny: Action Denial in Multimodal Large Language Models

• URL: http://arxiv.org/abs/2606.31187v1
• Date: Tue, 30 Jun 2026 06:16:17 GMT
• Title: Learning to Deny: Action Denial in Multimodal Large Language Models
• Abstract summary: Multimodal large language models (MLLMs) have rapidly advanced video understanding, achieving strong zero-shot and few-shot recognition across standard benchmarks.<n>Yet their ability to deny an action by recognizing when an activity is not happening despite strong contextual cues remains largely unexplored.<n>We introduce UCF101-AD, a benchmark consisting of paired Action-Presence and Action-Denial clips, designed to evaluate this capacity for denial.
• Score: 12.607064466360216
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Multimodal large language models (MLLMs) have rapidly advanced video understanding, achieving strong zero-shot and few-shot recognition across standard benchmarks. Yet their ability to deny an action by recognizing when an activity is not happening despite strong contextual cues remains largely unexplored. We introduce UCF101-AD, a large-scale benchmark consisting of paired Action-Presence and Action-Denial clips, designed to evaluate this capacity for denial. Each negative video in UCF101-AD preserves the same contextual and motion cues, including persons, objects, and locations, as its positive counterpart, but the defining action itself is explicitly absent. Evaluating 20 state-of-the-art MLLMs reveals a consistent failure: models that exceed 85% accuracy on the positive action classes collapse below 50% on their action-denial counterparts, indicating a strong inclination to affirm plausible actions rather than verify that they truly occur. This exposes a critical blind spot in modern video understanding: the inability to reason causally about whether a motion actually happens. To probe this issue, we explore a causal graph formulation, CausalAct, which expresses scene structure through natural-language prompts linking context, interaction, and motion. Incorporating such causal cues substantially reduces false positives, demonstrating that denial is a learnable reasoning skill. UCF101-AD provides a new lens for diagnosing and improving causal reasoning in multimodal models. Dataset and relevant code: https://github.com/raiyaan-abdullah/Learn-to-Deny.

Related papers

• ActiveScope: Actively Seeking and Correcting Perception for MLLMs [73.84325380140547]
  Multimodal Large Language Models (MLLMs) have demonstrated impressive vision-correction understanding, yet still struggle with fine-grained perception in high-resolution images.<n>We propose ActiveScope, a training-free framework that enhances MLLMs by actively seeking and correcting perception.
  arXiv  Detail & Related papers  (2026-06-23T08:17:15Z)
• How Far Are Video Models from True Multimodal Reasoning? [32.10821745418955]
  CLVG-Bench is an evaluation framework designed to probe video models' zero-shot reasoning capabilities.<n>It comprises more than 1,000 high-quality, manually annotated metadata across 6 categories and 47 subcategories.<n>We propose an Adaptive Video Evaluator (AVE) that aligns with human expert perception using minimal annotations.
  arXiv  Detail & Related papers  (2026-04-21T08:04:02Z)
• Vision-Language Models Unlock Task-Centric Latent Actions [75.53481518882275]
  We propose to utilize the common-sense reasoning abilities of Vision-Language Models (VLMs) to provide promptable representations.<n>We show that simply asking VLMs to ignore distractors can substantially improve latent action quality, yielding up to a six-fold increase in downstream success rates on Distracting MetaWorld.
  arXiv  Detail & Related papers  (2026-01-30T08:38:59Z)
• The Imperfective Paradox in Large Language Models [19.058068907991277]
  We investigate the Imperfective Paradox, where the past progressive aspect entails event realization for activities but not for accomplishments.<n>We introduce ImperfectiveNLI, a diagnostic dataset designed to probe this distinction across diverse semantic classes.<n>We uncover a pervasive Teleological Bias: models systematically hallucinate completion for goal-oriented events, often overriding explicit textual negation.
  arXiv  Detail & Related papers  (2026-01-14T10:57:16Z)
• Analyzing Reasoning Consistency in Large Multimodal Models under Cross-Modal Conflicts [74.47786985522762]
  We identify a critical failure mode termed textual inertia, where models tend to blindly adhere to the erroneous text while neglecting conflicting visual evidence.<n>We propose the LogicGraph Perturbation Protocol that structurally injects perturbations into the reasoning chains of diverse LMMs.<n>Results reveal that models successfully self-correct in less than 10% of cases and predominantly succumb to blind textual error propagation.
  arXiv  Detail & Related papers  (2026-01-07T16:39:34Z)
• When Alignment Fails: Multimodal Adversarial Attacks on Vision-Language-Action Models [75.16145284285456]
  We introduce VLA-Fool, a comprehensive study of multimodal adversarial robustness in embodied VLA models under both white-box and black-box settings.<n>We develop the first automatically crafted and semantically guided prompting framework.<n> Experiments on the LIBERO benchmark reveal that even minor multimodal perturbations can cause significant behavioral deviations.
  arXiv  Detail & Related papers  (2025-11-20T10:14:32Z)
• Disentangled Concepts Speak Louder Than Words:Explainable Video Action Recognition [22.38060746037401]
  We propose Disentangled Action aNd Context concept-based Explainable (DANCE) video action recognition.<n>DANCE predicts actions through disentangled concept types: motion dynamics, objects, and scenes.<n> Experiments on four datasets demonstrate that DANCE significantly improves explanation clarity with competitive performance.
  arXiv  Detail & Related papers  (2025-11-05T18:59:35Z)
• Enhancing Spatio-Temporal Zero-shot Action Recognition with Language-driven Description Attributes [54.50887214639301]
  We propose an innovative approach that harnesses web-crawled descriptions, leveraging a large-language model to extract relevant keywords.<n>This method reduces the need for human annotators and eliminates the laborious manual process of attribute data creation.<n>In our zero-shot experiments, our model achieves accuracies of 81.0%, 53.1%, and 68.9% on UCF-101, HMDB-51, and Kinetics-600, respectively.
  arXiv  Detail & Related papers  (2025-10-31T07:45:44Z)
• LLaVAction: evaluating and training multi-modal large language models for action recognition [46.473599879244716]
  We focus on evaluating and improving MLLMs to perform action recognition.<n>We reformulate EPIC-KITCHENS-100, one of the largest and most challenging egocentric action datasets.<n>We show that when we sample difficult incorrect answers as distractors, leading MLLMs struggle to recognize the correct actions.
  arXiv  Detail & Related papers  (2025-03-24T14:24:17Z)
• Hawk: Learning to Understand Open-World Video Anomalies [76.9631436818573]
  Video Anomaly Detection (VAD) systems can autonomously monitor and identify disturbances, reducing the need for manual labor and associated costs. We introduce Hawk, a novel framework that leverages interactive large Visual Language Models (VLM) to interpret video anomalies precisely. We have annotated over 8,000 anomaly videos with language descriptions, enabling effective training across diverse open-world scenarios, and also created 8,000 question-answering pairs for users' open-world questions.
  arXiv  Detail & Related papers  (2024-05-27T07:08:58Z)

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.