JarvisEvo: 向自进化式照片编辑代理迈进：协同编辑器-评估器优化
JarvisEvo: Towards a Self-Evolving Photo Editing Agent with Synergistic Editor-Evaluator Optimization
AI总结：
JarvisEvo通过协同编辑器-评估器优化，实现自进化式照片编辑代理，提升编辑质量和内容保真度。
摘要：
基于代理的编辑模型已显著提升了交互体验、处理质量和创作灵活性。然而，仍存在两个关键挑战：（1）指令幻觉，仅文本的链式思考（CoT）推理无法完全防止事实错误，因为存在固有的信息瓶颈；（2）奖励黑客，动态策略优化对抗静态奖励模型使代理能够利用奖励函数的缺陷。为了解决这些问题，我们提出了JarvisEvo，一个统一的图像编辑代理，通过迭代编辑、选择合适工具、评估结果和反思自身决策来模拟专家人类设计师，以优化输出结果。JarvisEvo提供了三个关键优势：（1）一种交错的多模态链式思考（iMCoT）推理机制，增强了指令遵循和编辑质量；（2）一种协同编辑器-评估器策略优化（SEPO）框架，使代理能够在没有外部奖励的情况下自我改进，有效缓解奖励黑客问题；（3）通过无缝整合Adobe Lightroom支持全局和局部精细编辑。在ArtEdit-Bench上，JarvisEvo在保真度编辑指标上平均比Nano-Banana高出18.95%，包括像素级内容保真度的显著提升44.96%。项目页面：https://jarvisevo.vercel.app/
英文摘要：
Agent-based editing models have substantially advanced interactive experiences, processing quality, and creative flexibility. However, two critical challenges persist: (1) instruction hallucination, text-only chain-of-thought (CoT) reasoning cannot fully prevent factual errors due to inherent information bottlenecks; (2) reward hacking, dynamic policy optimization against static reward models allows agents to exploit flaws in reward functions. To address these issues, we propose JarvisEvo, a unified image editing agent that emulates an expert human designer by iteratively editing, selecting appropriate tools, evaluating results, and reflecting on its own decisions to refine outcomes. JarvisEvo offers three key advantages: (1) an interleaved multimodal chain-of-thought (iMCoT) reasoning mechanism that enhances instruction following and editing quality; (2) a synergistic editor-evaluator policy optimization (SEPO) framework that enables self-improvement without external rewards, effectively mitigating reward hacking; and (3) support for both global and local fine-grained editing through seamless integration of Adobe Lightroom. On ArtEdit-Bench, JarvisEvo outperforms Nano-Banana by an average of 18.95% on preservative editing metrics, including a substantial 44.96% improvement in pixel-level content fidelity. Project page: https://jarvisevo.vercel.app/
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
