Linguistic Relative Policy Optimization for Video Anomaly Reasoning
2026-07-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors propose a new method called Linguistic Relative Policy Optimization (LRPO) to improve video anomaly detection without needing lots of labeled data or expert guidance. Their method uses language to capture and apply knowledge about anomalies from different scenarios, helping the model adapt without changing its parameters. They also introduce a reward system to align the model's decisions with human ideas of risk over time. Tests show their approach works better than previous methods without additional fine-tuning.
Video Anomaly DetectionMultimodal Large Language ModelsPolicy OptimizationSemantic RepresentationTrajectory OptimizationContext AdaptationReward FunctionTemporal ReasoningXD-ViolenceUCF-Crime
Authors
Jiaxu Leng, Jiankang Zheng, Mengjingcheng Mo, Zhanjie Wu, Haosheng Chen, Ji Gan, Xinbo Gao
Abstract
Video anomaly detection (VAD) with multimodal large language models has shown strong potential, yet most existing methods still depend on large-scale annotations or expert-designed priors, limiting their ability to acquire anomaly knowledge with as little human intervention as possible. To address this, we propose Linguistic Relative Policy Optimization (LRPO), which distills group-relative semantic advantages from multiple reasoning trajectories into a linguistically expressed anomaly experience prior, and adapts the model by injecting this prior into the context to steer its output distribution without any parameter updates. LRPO builds two complementary experience representations: general experience captures transferable anomaly preferences across scenarios, while scenario experience models context-dependent anomaly rules for targeted refinement. To further improve the learned experience, we introduce an anomaly alignment reward that guides trajectory optimization to match human risk preferences and reinforce temporally grounded reasoning. Extensive experiments on XD-Violence, UCF-Crime, and UBnormal demonstrate that LRPO significantly outperforms existing state-of-the-art methods under tuning-free settings.