mlsanityData to 7 Oct 2026

Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Models

arXiv 2602.22227 ↗First posted 24 Jan 2026cs.LG cs.AIcited by 0

On the map: Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness

Abstract (arXiv)

Despite their impressive capabilities, Multimodal Large Language Models (MLLMs) exhibit perceptual fragility when confronted with visually complex scenes. This weakness stems from a reliance on finite training datasets, which are prohibitively expensive to scale and impose a ceiling on model robustness. We introduce AOT-SFT, a large-scale adversarial dataset for bootstrapping MLLM robustness. Building on this, we propose AOT (Adversarial Opponent Training), a self-play framework that forges MLLM robustness by creating its own training data. Our method orchestrates a co-evolution between an image-editing Attacker and a Defender MLLM, where the Attacker generates a diverse and dynamic curriculum of image manipulations, forcing the Defender to adapt and improve. Extensive experiments demonstrate that AOT enhances the Defender's perceptual robustness and reduces hallucinations, establishing a scalable paradigm for training more reliable MLLMs.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 48 of its 58 references that are in the map. First posted on arXiv 24 Jan 2026.

Builds on

48 of 58 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Secure Tug-of-War (SecTOW): Iterative Defense-Attack Training with Reinforcement Learning for Multimodal Model Security29 Jul

Other references · 47

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  2. 2025Qwen3 Technical Reportcited by 8,195
  3. 2023Visual Instruction Tuningcited by 7,792
  4. 2025Qwen2.5-VL Technical Reportcited by 5,418
  5. 2024GPT-4o System Cardcited by 4,467
  6. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,655
  7. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
  8. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  9. 2016Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotationscited by 2,646
  10. 2023Code Llama: Open Foundation Models for Codecited by 2,105
  11. 2023MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGIcited by 2,044
  12. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,769
35 more
  1. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,696
  2. 2025Gemma 3 Technical Reportcited by 1,686
  3. 2021Ego4D: Around the World in 3,000 Hours of Egocentric Videocited by 1,530
  4. 2024Are We on the Right Way for Evaluating Large Vision-Language Models?cited by 893
  5. 2016A Diagram Is Worth A Dozen Imagescited by 793
  6. 2023WizardCoder: Empowering Code Large Language Models with Evol-Instructcited by 728
  7. 2024SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilitiescited by 679
  8. 2025Flow-GRPO: Training Flow Matching Models via Online RLcited by 605
  9. 2023HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Modelscited by 572
  10. 2024BLINK: Multimodal Large Language Models Can See but Not Perceivecited by 529
  11. 2020Overfitting in adversarially robust deep learningcited by 504
  12. 2023V*: Guided Visual Search as a Core Mechanism in Multimodal LLMscited by 459
  13. 2025Absolute Zero: Reinforced Self-play Reasoning with Zero Datacited by 323
  14. 2023Magicoder: Empowering Code Generation with OSS-Instructcited by 251
  15. 2023Multimodal Large Language Models: A Surveycited by 193
  16. 2024LLaVA-UHD: an LMM Perceiving Any Aspect Ratio and High-Resolution Imagescited by 193
  17. 2024Divide, Conquer and Combine: A Training-Free Framework for High-Resolution Image Perception in Multimodal Large Language Modelscited by 151
  18. 2022Foundations and Trends in Multimodal Machine Learning: Principles, Challenges, and Open Questionscited by 116
  19. 2023RLTF: Reinforcement Learning from Unit Test Feedbackcited by 110
  20. 2025Why Is Spatial Reasoning Hard for VLMs? An Attention Mechanism Perspective on Focus Areascited by 93
  21. 2024Beyond LLaVA-HD: Diving into High-Resolution Large Multimodal Modelscited by 75
  22. 2025Self-Challenging Language Model Agentscited by 56
  23. 2025Co-Evolving LLM Coder and Unit Tester via Reinforcement Learningcited by 51
  24. 2025SeRL: Self-Play Reinforcement Learning for Large Language Models with Limited Datacited by 50
  25. 2025Vision-Zero: Scalable VLM Self-Improvement via Strategic Gamified Self-Playcited by 37
  26. 2025Learning to Solve and Verify: A Self-Play Framework for Code and Test Generationcited by 34
  27. 2025Self-Improving Language Models for Evolutionary Program Synthesis: A Case Study on ARC-AGIcited by 34
  28. 2025G1: Bootstrapping Perception and Reasoning Abilities of Vision-Language Model via Reinforcement Learningcited by 29
  29. 2024Dynamic Multimodal Evaluation with Flexible Complexity by Vision-Language Bootstrappingcited by 22
  30. 2025OneReward: Unified Mask-Guided Image Generation via Multi-Task Human Preference Learningcited by 22
  31. 2024Survey of Large Multimodal Model Datasets, Application Categories and Taxonomycited by 17
  32. 2025Analysing the Robustness of Vision-Language-Models to Common Corruptionscited by 16
  33. 2025Addressing Benchmarking Gaps in Large Language Models for Health and Medicine with Dynamic Red-Teamingcited by 10
  34. 2025On the robustness of multimodal language model towards distractionscited by 9
  35. 2025Vision-Language Model Dialog Games for Self-Improvementcited by 4

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Investigating Adversarial Robustness of Multi-modal Large Language Models
  2. 2025Survey of Adversarial Robustness in Multimodal Large Language Models
  3. 2026Adversarial Attacks on Multimodal Large Language Models: A Comprehensive Survey
  4. 2023On Evaluating Adversarial Robustness of Large Vision-Language Models
  5. 2025Robust-LLaVA: On the Effectiveness of Large-Scale Robust Image Encoders for Multi-modal Large Language Models
  6. 2026Exploring Adversarial Robustness and Safety Alignment in Multilingual Multi-Modal Large Language Models
  7. 2024Multimodal Adversarial Defense for Vision-Language Models by Leveraging One-To-Many Relationships
  8. 2025SmoothGuard: Defending Multimodal Large Language Models with Noise Perturbation and Clustering Aggregation