Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.17969 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909053970546688 |
|---|---|
| author | Liu, Jinming Feng, Ruoyu Wang, Yuqi Zeng, Wenjun Jin, Xin |
| author_facet | Liu, Jinming Feng, Ruoyu Wang, Yuqi Zeng, Wenjun Jin, Xin |
| contents | Despite rapid advances in text-to-image generation, faithfully realizing user intent remains challenging, often requiring manual multi-turn trial and error. To automate this process, existing systems rely on either simple prompt rewriting or closed-loop agents driven by hand-crafted rules, rather than learning to adapt actions to the evolving generation process. In this paper, we reformulate image generation as a state-conditioned action-making problem and propose Generation Navigator, a multi-turn T2I agent that learns to dynamically steer the generation trajectory and output the next action. However, training this agent via reinforcement learning introduces a critical credit assignment challenge: naively rewarding a trajectory based solely on a single state assigns equal credit to all actions in the rollout, ignores the quality dynamics across turns, and fails to distinguish actions that improve the trajectory from those that degrade it or waste turns without progress. We resolve this with PRE-GRPO (Peak-Retention-Efficiency Group Relative Policy Optimization), a trajectory-level reinforcement learning objective that explicitly rewards discovering a high-quality image (Peak), avoiding subsequent quality degradation across turns (Retention), and minimizing unnecessary turns (Efficiency). Experiments show substantial improvements across benchmarks, reaching a WISE score of 0.90 and 79.06% reasoning accuracy on T2I-ReasonBench. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_17969 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Generation Navigator: A State-Aware Agentic Framework for Image Generation Liu, Jinming Feng, Ruoyu Wang, Yuqi Zeng, Wenjun Jin, Xin Computer Vision and Pattern Recognition Despite rapid advances in text-to-image generation, faithfully realizing user intent remains challenging, often requiring manual multi-turn trial and error. To automate this process, existing systems rely on either simple prompt rewriting or closed-loop agents driven by hand-crafted rules, rather than learning to adapt actions to the evolving generation process. In this paper, we reformulate image generation as a state-conditioned action-making problem and propose Generation Navigator, a multi-turn T2I agent that learns to dynamically steer the generation trajectory and output the next action. However, training this agent via reinforcement learning introduces a critical credit assignment challenge: naively rewarding a trajectory based solely on a single state assigns equal credit to all actions in the rollout, ignores the quality dynamics across turns, and fails to distinguish actions that improve the trajectory from those that degrade it or waste turns without progress. We resolve this with PRE-GRPO (Peak-Retention-Efficiency Group Relative Policy Optimization), a trajectory-level reinforcement learning objective that explicitly rewards discovering a high-quality image (Peak), avoiding subsequent quality degradation across turns (Retention), and minimizing unnecessary turns (Efficiency). Experiments show substantial improvements across benchmarks, reaching a WISE score of 0.90 and 79.06% reasoning accuracy on T2I-ReasonBench. |
| title | Generation Navigator: A State-Aware Agentic Framework for Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.17969 |