DVR: Dynamic visual re-reading for multimodal large language models via sub-question feedback
Authors:
- Yuyao ZhangYZYuyao Zhang
- Zefeng ChenZCZefeng Chen
- Haizhai ZhangHZHaizhai Zhang
- Guorui ChenGCGuorui Chen
- Chang YangCYChang Yang
- Zhijiang LiZhijiang Li
- 10 July 2026
- journal article
- Published by Elsevier in Neurocomputing
- Vol. 700
Abstract
No abstract availableThis publication has 8 references indexed in Scilit:
- Accelerating Pre-training of Multimodal LLMs via Chain-of-Sight2024
- Are We on the Right Way for Evaluating Large Vision-Language Models?2024
- Exploring Question Decomposition for Zero-Shot VQA2023
- Visual Instruction Tuning2023
- InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning2023
- Mass-Producing Failures of Multimodal Systems with Language Models2023
- Optimal eye movement strategies in visual searchNature, 2005
- Why a Diagram is (Sometimes) Worth Ten Thousand WordsCognitive Science, 1987