# Controllable Multi-label Video Safety Detection via Adaptive Tversky Policy Optimization

arXiv: https://arxiv.org/abs/2610.02019 · HTML: https://mlsanity.com/paper/2610.02019/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 33 of its 49 references that are in the map.
- First posted on arXiv 1 Oct 2026.

## Builds on (most cited first, 30 of 33)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [Video-ChatGPT: Towards Detailed Video Understanding via Large Vision and Language Models](https://mlsanity.com/paper/2306.05424/) (2023)
- [Real-world Anomaly Detection in Surveillance Videos](https://mlsanity.com/paper/1801.04264/) (2018)
- [Video-R1: Reinforcing Video Reasoning in MLLMs](https://mlsanity.com/paper/2503.21776/) (2025)
- [Future Frame Prediction for Anomaly Detection -- A New Baseline](https://mlsanity.com/paper/1712.09867/) (2017)
- [Learning Memory-guided Normality for Anomaly Detection](https://mlsanity.com/paper/2003.13228/) (2020)
- [VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning](https://mlsanity.com/paper/2504.06958/) (2025)
- [Not only Look, but also Listen: Learning Multimodal Violence Detection under Weak Supervision](https://mlsanity.com/paper/2007.04687/) (2020)
- [Improve Vision Language Model Chain-of-thought Reasoning](https://mlsanity.com/paper/2410.16198/) (2024)
- [SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generation](https://mlsanity.com/paper/2410.12761/) (2024)
- [Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity](https://mlsanity.com/paper/2412.06171/) (2024)
- [T2VSafetyBench: Evaluating the Safety of Text-to-Video Generative Models](https://mlsanity.com/paper/2407.05965/) (2024)
- [VideoChat-R1.5: Visual Test-Time Scaling to Reinforce Multimodal Reasoning by Iterative Perception](https://mlsanity.com/paper/2509.21100/) (2025)
- [Holmes-VAD: Towards Unbiased and Explainable Video Anomaly Detection via Multi-modal LLM](https://mlsanity.com/paper/2406.12235/) (2024)
- [Hawk: Learning to Understand Open-World Video Anomalies](https://mlsanity.com/paper/2405.16886/) (2024)
- [GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning](https://mlsanity.com/paper/2505.11049/) (2025)
- [SafeWatch: An Efficient Safety-Policy Following Video Guardrail Model with Transparent Explanations](https://mlsanity.com/paper/2412.06878/) (2024)
- [Reinforcement Learning Tuning for VideoLLMs: Reward Design and Data Efficiency](https://mlsanity.com/paper/2506.01908/) (2025)
- [Vad-R1: Towards Video Anomaly Reasoning via Perception-to-Cognition Chain-of-Thought](https://mlsanity.com/paper/2505.19877/) (2025)
- [TempSamp-R1: Effective Temporal Sampling with Reinforcement Fine-Tuning for Video LLMs](https://mlsanity.com/paper/2509.18056/) (2025)
- [Anomize: Better Open Vocabulary Video Anomaly Detection](https://mlsanity.com/paper/2503.18094/) (2025)
- [T2Vs Meet VLMs: A Scalable Multimodal Dataset for Visual Harmfulness Recognition](https://mlsanity.com/paper/2409.19734/) (2024)
- [SEA: Low-Resource Safety Alignment for Multimodal Large Language Models via Synthetic Embeddings](https://mlsanity.com/paper/2502.12562/) (2025)
- [SafeVid: Toward Safety Aligned Video Large Multimodal Models](https://mlsanity.com/paper/2505.11926/) (2025)
- [ExPO-HM: Learning to Explain-then-Detect for Hateful Meme Detection](https://mlsanity.com/paper/2510.08630/) (2025)

## Abstract (arXiv)

The rapid growth of video-based social media has increased users' exposure to harmful content, creating a need for reliable automated video safety detection. Although recent Vision-Language Models (VLMs) show strong video understanding capabilities, existing harmful video detection systems face two key limitations: they typically reduce safety detection to binary classification, overlooking the inherently multi-label nature of unsafe videos, and they rely on static training objectives that do not support controllable precision-recall trade-offs, though the desired operating point may vary across moderation pipelines and unsafe categories. To address these gaps, we propose Adaptive Tversky Policy Optimization (ATPO), a reinforcement learning framework for Multi-label Video Safety Detection (Multi-VSD). ATPO introduces the Adaptive Tversky Reward (ATR), which dynamically adjusts false-positive and false-negative penalties during training to enable controllable precision-recall trade-offs. Experiments on SafeWatch-Bench and XD-Violence show that ATPO substantially improves multi-label performance, increasing the Jaccard Index from 40.66 to 75.44 on SafeWatch-Bench-Real. Moreover, ATR enables reliable steering of the precision-recall operating point, supporting deployment scenarios with heterogeneous policy requirements. Code and checkpoints are provided at https://bruceyg.github.io/ATPO-project-page/ .