2025/11/24 by Wang, Qiang, Gao, Xinyuan, Dong, SongLin +4 · 1 citation
Computer Science · #Artificial Intelligence (cs.AI) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis #Machine Learning (cs.LG) #Multimedia (cs.MM) #Multimodal Machine Learning Applications #Topic Modeling
paper · doi:10.48550/arxiv.2511.19436
openalex publication_date 2025/11/24 · openalex created_date 2025/11/27 · openalex updated_date 2026/07/28
We present VDC-Agent, a self-evolving framework for Video Detailed Captioning that requires neither human annotations nor larger teacher models. The agent forms a closed loop of caption generation, principle-guided scoring (score and textual suggestions), and prompt refinement. When caption quality regresses, a self-reflection path leverages the previous chain-of-thought to amend the update. Running this process on unlabeled videos produces trajectories of (caption, score) pairs. We convert the trajectories into preference tuples and filter out samples with JSON parsing errors, resulting in VDC-Agent-19K, which contains 18,886 automatically constructed pairs. We then fine-tune the base MLLM on this dataset using an easy-to-hard curriculum direct preference optimization. Built on Qwen2.5-VL-7B-Instruct, our VDC-Agent-7B attains state-of-the-art performance on the VDC benchmark with 49.08% average accuracy and 2.50 score, surpassing specialized video captioners and improving over the base model by +5.13% accuracy and +0.27 score at similar inference cost.