mlsanityData to 14 Oct 2026

Parser-Free VLM Verification for Federated Weakly Supervised Video Anomaly Detection

arXiv 2609.07455 ↗First posted 7 Sep 2026cs.CV cs.AI cs.LGcited by 0

On the map: Multimodal LLMs / vision-language models → Chart understanding and anomaly detection

Abstract (arXiv)

How can vision-language models help video anomaly detection (VAD) when surveillance data remain distributed, weakly labeled, and resource-constrained? Most weakly supervised VAD methods assume centralized training; recent VLM-based extensions further rely on dense inference, generated explanations, or additional adaptation. We introduce a lightweight federated MIL-VLM cascade in which only a compact MIL scorer is trained across clients, while a frozen VLM verifies high-scoring suspect segments post hoc. We study two VLM feedback interfaces: parsed text-generation decisions and a logit-based interface that extracts a continuous anomaly score from next-token Yes/No probabilities. Experiments on UCF-Crime with InternVL3.5-2B and Qwen3-VL-2B-Instruct show that text-generation verification can improve frame-level AUC after diagnostic temporal post-processing, but remains sensitive to prompts, parsers, model choice, and smoothing. In contrast, the logit interface provides a fixed parser-free signal that improves both frame-level AUC and frame-level AP over the MIL baseline across both VLMs, without temporal post-processing in its main configuration. Since suspect segments are updated independently once available, next-token logit feedback provides a simple segment-local alternative to text-generation verification.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Chart understanding and anomaly detection, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 15 of its 19 references that are in the map. First posted on arXiv 7 Sep 2026.

Builds on

15 of 19 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025Vad-R1: Towards Video Anomaly Reasoning via Perception-to-Cognition Chain-of-Thought26 May
  2. 2025SlowFastVAD: Video Anomaly Detection via Integrating Simple Detector and RAG-Enhanced Vision-Language Model14 Apr
  3. 2024Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity9 Dec
  4. 2024VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models2 Dec
  5. 2023VadCLIP: Adapting Vision-Language Models for Weakly Supervised Video Anomaly Detection22 Aug

Other references · 10

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,669
  3. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,510
  4. 2018Real-world Anomaly Detection in Surveillance Videoscited by 577
  5. 2021Weakly-supervised Video Anomaly Detection with Robust Temporal Feature Magnitude Learningcited by 230
  6. 2023Dual Memory Units with Uncertainty Regulation for Weakly Supervised Video Anomaly Detectioncited by 90
  7. 2022MGFN: Magnitude-Contrastive Glance-and-Focus Network for Weakly-Supervised Video Anomaly Detectioncited by 89
  8. 2024Harnessing Large Language Models for Training-free Video Anomaly Detectioncited by 87
  9. 2023Unbiased Multiple Instance Learning for Weakly Supervised Video Anomaly Detectioncited by 64
  10. 2024Follow the Rules: Reasoning for Video Anomaly Detection with Large Language Modelscited by 57

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026CoReVAD: A Contextual Reasoning Framework for Training-Free Video Anomaly Detection
  2. 2026Benchmarking Compact VLMs for Clip-Level Surveillance Anomaly Detection Under Weak Supervision
  3. 2026Anomaly-Guided Observation and Two-Level Semantic Feedback for Training-Free Video Anomaly Analysis
  4. 2024Video Anomaly Detection in 10 Years: A Survey and Outlook
  5. 2026MuST-VAD: Mutual Structured Learning for Video Anomaly Detection
  6. 2024Quo Vadis, Anomaly Detection? LLMs and VLMs in the Spotlight
  7. 2026COPRA: Conditional Parameter Adaptation with Reinforcement Learning for Video Anomaly Detection
  8. 2026Context-structured Video Anomaly Detection with Large Vision-Language Models