UniGame：将统一多模态模型转化为其自身对抗者

摘要

统一多模态模型（UMMs）通过单一架构在理解与生成任务中均展现出卓越性能。然而，此类模型仍存在根本性矛盾：理解任务偏好紧凑的嵌入表示，而生成任务则需要重构丰富的表征。这种结构性权衡导致决策边界失准、跨模态连贯性下降，并在分布偏移和对抗性攻击下表现出更高的脆弱性。本文提出UniGame——一种直接针对上述不一致性的自对抗后训练框架。通过在共享令牌接口施加轻量化扰动器，该框架使生成分支能够主动探寻并挑战脆弱的理解环节，让模型自身成为其对抗者。实验表明，UniGame显著提升模型一致性（+4.6%），同时在理解能力（+3.6%）、生成质量（+0.02）以及分布外鲁棒性（NaturalBench和AdVQA数据集分别提升+4.8%和+6.2%）方面实现显著进步。该框架与架构无关，仅引入不足1%的额外参数，且可与现有后训练方法互补。这些成果表明，对抗性自我博弈是提升未来多模态基础模型连贯性、稳定性与统一能力的普适性有效原则。项目代码已开源：https://github.com/AIFrontierLab/UniGame

English

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), 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/UniGame

UniGame：将统一多模态模型转化为其自身对抗者

UniGame: Turning a Unified Multimodal Model Into Its Own Adversary

摘要

Support