Zhijiang Tang (唐治江)

Zhijiang Tang (唐治江)

M.S. student in AI at UCAS (Hangzhou Institute for Advanced Study). LLM post-training, multimodal evaluation, and agent RL.

CCCaption: Dual-Reward RL for Complete and Correct Image Captioning

Published in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026

First author. Accepted to CVPR 2026 (CCF-A). Proposes completeness and correctness dual rewards with dynamic query sampling for efficient RL post-training; CCCaption-2B achieves capability comparable to Qwen3-VL-32B.

Recommended citation: Tang, Z., Wang, L., Qi, J., Jiang, W., Hou, P., Zeng, A., & Huang, J. (2026). "CCCaption: Dual-Reward RL for Complete and Correct Image Captioning." CVPR 2026.
Download Paper