Mind-Brush: Integrating Agentic Cognitive Search and Reasoning into Image Generation

Jun He, Junyan Ye, Zilong Huang, Dongzhi Jiang, Chenjue Zhang, Leqi Zhu, Renrui Zhang, Xiang Zhang, Weijia Li · 2026

Read Mind-Brush: Integrating Agentic Cognitive Search and Reasoning into Image Generation by Jun He, Junyan Ye, Zilong Huang, Dongzhi Jiang, Chenjue Zhang, Leqi Zhu, Renrui Zhang, Xiang Zhang, Weijia Li (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-01756.txt.

Description

Jun He and colleagues turn text to image generation from pixel decoding into a knowledge driven workflow. Static decoders miss intentions, unified models improve intent but struggle with knowledge reasoning, and fixed priors cannot track change. Their Mind-Brush simulates a think research create loop retrieving evidence that grounds unseen concepts and using tools to resolve constraints. They propose Mind-Bench with 500 samples spanning news, concepts, math and Geo-Reasoning. Mind-Brush improves models, leaping from zero to one for Qwen-Image on Mind-Bench and beating WISE and RISE.

Catalog record

CREATORJun He, Junyan Ye, Zilong Huang, Dongzhi Jiang, Chenjue Zhang, Leqi Zhu, Renrui Zhang, Xiang Zhang, Weijia Li
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54834
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01756

Browse the full library · About the Tower of Babel