arXiv · cs.AI · Jinmei Liu, Haoru Li, Zhenhong Sun, Chaofeng Chen, Yatao Bian, Bo Wang, Daoyi Dong, Chunlin Chen, Zhi Wang · 5 小时前模型评分44
DRIFT：缓解强化微调中多样性坍缩的通用图像生成框架
Beyond the Dirac Delta: Mitigating Diversity Collapse in Reinforcement Fine-Tuning for Versatile Image Generation
导读
研究者提出 DRIFT 框架，在强化微调过程中持续激励输出多样性，以缓解策略坍缩为狄拉克 delta 分布的多样性坍缩问题。该框架从采样、提示词和优化三方面入手，在任务对齐与生成多样性上取得明显帕累托优势：对齐相同时多样性提升 3.88%~81.42%，多样性相同时对齐提升 14.38%~56.45%。代码已公开。
分类置信度
中 · 待人工复核
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
事件类型：研究发布
证据摘录
- “To address this challenge, we propose \textbf{DRIFT} (\textbf{D}ive\textbf{R}sity-\textbf{I}ncentivized Reinforcement \textbf{F}ine-\textbf{T}uning for Versatile Image Generation), an innovative framework that systematically incentivizes output diversity throughout the on-policy fine-tuning process, reconciling strong task alignment with high generation diversity to enhance versatility essential for applications that demand diverse candidate generations.”
- “Our code is available at \textcolor{magenta}{\href{ this https URL }{ this https URL }}.”
来源：arXiv · cs.AI · arxiv.org