GenPilot: A Multi-Agent System for Test-Time Prompt Optimization in Image Generation
Published in Findings of the Association for Computational Linguistics: EMNLP 2025, 2025
Text-to-image models often misinterpret complex and lengthy prompts, resulting in semantic inconsistencies and missing details. GenPilot is a flexible, model-agnostic test-time prompt optimization strategy that operates directly on the input text, integrating error analysis, clustering-based adaptive exploration and fine-grained verification.
ACL Anthology | DOI: 10.18653/v1/2025.findings-emnlp.49
Recommended citation: Wen Ye, Zhaocheng Liu, Yuwei Gui, Tingyu Yuan, Yunyue Su, Bowen Fang, Chaoyang Zhao, Qiang Liu, Liang Wang. (2025). "GenPilot: A Multi-Agent System for Test-Time Prompt Optimization in Image Generation." Findings of the Association for Computational Linguistics: EMNLP 2025, pp. 929-958.
Download Paper