# MM-VeriAgent: Learning to Use Extensive Tools to Verify Multimodal Misinformation with Reinforcement Learning

arXiv: https://arxiv.org/abs/2609.30698 · HTML: https://mlsanity.com/paper/2609.30698/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 20 of its 34 references that are in the map.
- First posted on arXiv 25 Sep 2026.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)

## Builds on (most cited first, 20 of 20)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning](https://mlsanity.com/paper/2503.09516/) (2025)
- [FEVER: a large-scale dataset for Fact Extraction and VERification](https://mlsanity.com/paper/1803.05355/) (2018)
- [CNN-generated images are surprisingly easy to spot... for now](https://mlsanity.com/paper/1912.11035/) (2019)
- [Towards Universal Fake Image Detectors that Generalize Across Generative Models](https://mlsanity.com/paper/2302.10174/) (2023)
- [ReTool: Reinforcement Learning for Strategic Tool Use in LLMs](https://mlsanity.com/paper/2504.11536/) (2025)
- [Image Manipulation Detection by Multi-View Multi-Scale Supervision](https://mlsanity.com/paper/2104.06832/) (2021)
- [Detecting and Grounding Multi-Modal Media Manipulation](https://mlsanity.com/paper/2304.02556/) (2023)
- [VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use](https://mlsanity.com/paper/2509.01055/) (2025)
- [GLiNER: Generalist Model for Named Entity Recognition using Bidirectional Transformer](https://mlsanity.com/paper/2311.08526/) (2023)
- [MMFakeBench: A Mixed-Source Multimodal Misinformation Detection Benchmark for LVLMs](https://mlsanity.com/paper/2406.08772/) (2024)
- [Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable](https://mlsanity.com/paper/2505.14359/) (2025)
- [Simulating Environments with Reasoning Models for Agent Training](https://mlsanity.com/paper/2511.01824/) (2025)
- [DEFAME: Dynamic Evidence-based FAct-checking with Multimodal Experts](https://mlsanity.com/paper/2412.10510/) (2024)
- [Faking Fake News for Real Fake News Detection: Propaganda-loaded Training Data Generation](https://mlsanity.com/paper/2203.05386/) (2022)
- [FKA-Owl: Advancing Multimodal Fake News Detection through Knowledge-Augmented LVLMs](https://mlsanity.com/paper/2403.01988/) (2024)
- [TVCACHE: A Stateful Tool-Value Cache for Post-Training LLM Agents](https://mlsanity.com/paper/2602.10986/) (2026)
- [CacheRL:Multi-Turn Tool-Calling Agents via Cached Rollouts and Hybrid Reward](https://mlsanity.com/paper/2606.14179/) (2026)

## Abstract (arXiv)

Real-world multimodal misinformation often involves mixed forgery sources, requiring sample-specific detection strategies. Existing tool-augmented methods rely on predefined workflows or inference-time planning, limiting adaptability or increasing inference cost. To address this issue, we introduce MM-VeriAgent, which learns to verify mixed-source multimodal misinformation with tools. We first build MM-VeriTools, a specialized toolkit for misinformation detection agents. By benchmarking various candidate models and methods on the sub-tasks required by mixed-source detection, we select the strongest for textual, visual, and cross-modal forgery analysis and encapsulate them as callable tools with a unified interface. On top of this toolkit, we train the LVLM agent with reinforcement learning to teach it how to use these tools to better solve mixed-source detection. Since many of the tools are specialized models whose online execution at every rollout severely limits RL efficiency, we further introduce Tool-Execution Cache, which pre-executes candidate tool calls and reuses their cached outputs during training. This preserves multi-step rollouts while reducing online tool execution, largely improving the training efficiency.Experiments on MMFakeBench demonstrate substantial accuracy gains over the base model without explicit tool search at inference time. Ablation and efficiency analyses further validate the learned tool-use policy and show that Tool-Execution Cache reduces online tool executions during training.