Dynamic Adversarial Reinforcement Learning for Robust Multimodal Large Language Models

TL;DR

This paper presents AOT, a self-play adversarial training framework that significantly improves the robustness and reliability of Multimodal Large Language Models against complex visual scenes.

Contribution

It introduces AOT-SFT, a large-scale adversarial dataset, and a self-play training method that co-evolves an attacker and defender to enhance MLLM robustness.

Findings

AOT improves perceptual robustness of MLLMs.

Reduces hallucinations in multimodal models.

Establishes scalable adversarial training paradigm.

Abstract

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 \textbf{AOT-SFT}, a large-scale adversarial dataset for bootstrapping MLLM robustness. Building on this, we propose \textbf{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.