Look Back to Move Forward: Learning to Revisit for Agentic Compositional Image Generation
In this work, we introduce GenRevisit, an agentic framework that extends action selection to joint continuation-state and action selection.
ProofPaper ↗
Key points
- Text-to-image models have made remarkable progress in visual fidelity and instruction following, yet complex compositional generation remains challenging.
- Existing agentic approaches address this limitation through iterative generation and editing, adapting their actions based on interaction history.
- We train GenRevisit with SFT followed by agentic reinforcement learning.
- To address the coarse credit assignment of final outcome reward in GRPO, which assigns the same trajectory-level advantage to all actions, we further propose Normalized Action Credit Assignment (NACA) to redistribute positive advantage according to each action's contribution to reducing the remaining objective.
Sources (1)
- [1]Look Back to Move Forward: Learning to Revisit for Agentic Compositional Image GenerationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 05:59 AM
In this work, we introduce GenRevisit, an agentic framework that extends action selection to joint continuation-state and action selection.
Text-to-image models have made remarkable progress in visual fidelity and instruction following, yet complex compositional generation remains challenging.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 12, 2026AUGUstus: Autonomous Driving VLA Model with Unified Generation and Understanding
- Oct 11, 2026AesCode: Aesthetic Code Generation with Decoupled Cross-Modal Rewards
- Oct 11, 2026RSIPO: Recursive Self-Improving Policy Optimization for Agentic Reinforcement Learning via Skill-Policy Co-Evolution
- Oct 10, 2026Structural Chain-of-Thought: Reasoning with Executable Computational Graphs
- Oct 10, 2026vllm-sr/d3
- Oct 9, 2026MENTOR: Learning from Failure with On-Policy Pedagogical Intervention