UniGame: Turning a Unified Multimodal Model Into Its Own Adversary

TL;DR

UniGame introduces a self-adversarial post-training method for unified multimodal models, significantly enhancing their understanding, generation, and robustness by actively challenging their own representations.

Contribution

It presents a lightweight, architecture-agnostic framework that improves UMMs through adversarial self-play, with less than 1% additional parameters and compatible with existing methods.

Findings

Improves consistency by +4.6% on GenEval

Enhances understanding by +3.6% and generation by +0.02

Boosts robustness by +4.8% and +6.2% on NaturalBench and AdVQA

Abstract

Unified Multimodal Models (UMMs) have shown impressive performance in both understanding and generation with a single architecture. However, UMMs still exhibit a fundamental inconsistency: understanding favors compact embeddings, whereas generation favors reconstruction-rich representations. This structural trade-off produces misaligned decision boundaries, degraded cross-modal coherence, and heightened vulnerability under distributional and adversarial shifts. In this paper, we present UniGame, a self-adversarial post-training framework that directly targets the inconsistencies. By applying a lightweight perturber at the shared token interface, UniGame enables the generation branch to actively seek and challenge fragile understanding, turning the model itself into its own adversary. Experiments demonstrate that UniGame significantly improves the consistency (+4.6%). Moreover, it also achieves substantial improvements in understanding (+3.6%), generation (+0.02)on GenEval, out-of-distribution and adversarial robustness (+4.8% and +6.2% on NaturalBench and AdVQA). The framework is architecture-agnostic, introduces less than 1% additional parameters, and is complementary to existing post-training methods. These results position adversarial self-play as a general and effective principle for enhancing the coherence, stability, and unified competence of future multimodal foundation models. The official code is available at: https://github.com/AIFrontierLab/TorchUMM