11 papers
Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping
Haoyuan Sun, Jing Wang, Yuxin Song +9
Recently, post-training methods based on reinforcement learning, with a particular focus on Group Relative Policy Optimization (GRPO), have emerged as the robust paradigm for furth…
When to Lock Attention: Training-Free KV Control in Video Diffusion
Tianyi Zeng, Jincheng Gao, Tianyi Wang +8
Maintaining background consistency while enhancing foreground quality remains a core challenge in video editing. Injecting full-image information often leads to background artifact…
Twin Co-Adaptive Dialogue for Progressive Image Generation
Jianhui Wang, Yangfan He, Yan Zhong +12
Modern text-to-image generation systems have enabled the creation of remarkably realistic and high-quality visuals, yet they often falter when handling the inherent ambiguities in…
Enhancing Intent Understanding for Ambiguous prompt: A Human-Machine Co-Adaption Strategy
Yangfan He, Jianhui Wang, Yijin Wang +18
Current image generation systems produce high-quality images but struggle with ambiguous user prompts, making interpretation of actual user intentions difficult. Many users must mo…
TDRI: Two-Phase Dialogue Refinement and Co-Adaptation for Interactive Image Generation
Yuheng Feng, Jianhui Wang, Kun Li +5
Although text-to-image generation technologies have made significant advancements, they still face challenges when dealing with ambiguous prompts and aligning outputs with user int…
OMR-Diffusion:Optimizing Multi-Round Enhanced Training in Diffusion Models for Improved Intent Understanding
Kun Li, Jianhui Wang, Miao Zhang +1
Generative AI has significantly advanced text-driven image generation, but it still faces challenges in producing outputs that consistently align with evolving user preferences and…