Papers by Chan-Wei Hu
Region-R1: Reinforcing Query-Side Region Cropping for Multi-Modal Re-Ranking (2026.findings-acl)
Copied to clipboard
| Challenge: | Multi-modal retrieval-augmented generation relies heavily on re-rankers to surface the most relevant evidence for image-question queries. |
| Approach: | They propose a query-side region cropping framework that makes region selection a decision-making problem during re-ranking. |
| Outcome: | The proposed framework learns to retain the full image or focus only on a question-relevant region before scoring the retrieved candidates. |
Re-Align: Aligning Vision Language Models via Retrieval-Augmented Direct Preference Optimization (2025.emnlp-main)
Copied to clipboard
Shuo Xing, Peiran Li, Yuping Wang, Ruizheng Bai, Yueqi Wang, Chan-Wei Hu, Chengxuan Qian, Huaxiu Yao, Zhengzhong Tu
| Challenge: | emergence of large Vision Language Models (VLMs) has broadened the capabilities of single-modal Large Language Model (LLM) but VLMs are prone to significant hallucinations, especially in the form of cross-modal inconsistencies. |
| Approach: | They propose a new alignment framework that leverages image retrieval to integrate both textual and visual preference signals. |
| Outcome: | The proposed framework mitigates hallucinations more effectively than previous methods . it maintains robustness and scalability across a wide range of VLM sizes and architectures . |