Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.14446 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914163082657792 |
|---|---|
| author | Gao, Hong Bao, Yiming Tu, Xuezhen Xu, Yutong Jin, Yue Mu, Yiyang Zhong, Bin Yue, Linan Zhang, Min-Ling |
| author_facet | Gao, Hong Bao, Yiming Tu, Xuezhen Xu, Yutong Jin, Yue Mu, Yiyang Zhong, Bin Yue, Linan Zhang, Min-Ling |
| contents | Video understanding requires not only visual recognition but also complex reasoning. While Vision-Language Models (VLMs) demonstrate impressive capabilities, they typically process videos largely in a single-pass manner with limited support for evidence revisit and iterative refinement. While recently emerging agent-based methods enable long-horizon reasoning, they either depend heavily on expensive proprietary models or require extensive agentic RL training. To overcome these limitations, we propose Agentic Video Intelligence (AVI), a flexible and training-free framework that can mirror human video comprehension through system-level design and optimization. AVI introduces three key innovations: (1) a human-inspired three-phase reasoning process (Retrieve-Perceive-Review) that ensures both sufficient global exploration and focused local analysis, (2) a structured video knowledge base organized through entity graphs, along with multi-granularity integrated tools, constituting the agent's interaction environment, and (3) an open-source model ensemble combining reasoning LLMs with lightweight base CV models and VLM, eliminating dependence on proprietary APIs or RL training. Experiments on LVBench, VideoMME-Long, LongVideoBench, and Charades-STA demonstrate that AVI achieves competitive performance while offering superior interpretability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_14446 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Agentic Video Intelligence: A Flexible Framework for Advanced Video Exploration and Understanding Gao, Hong Bao, Yiming Tu, Xuezhen Xu, Yutong Jin, Yue Mu, Yiyang Zhong, Bin Yue, Linan Zhang, Min-Ling Computer Vision and Pattern Recognition Artificial Intelligence Video understanding requires not only visual recognition but also complex reasoning. While Vision-Language Models (VLMs) demonstrate impressive capabilities, they typically process videos largely in a single-pass manner with limited support for evidence revisit and iterative refinement. While recently emerging agent-based methods enable long-horizon reasoning, they either depend heavily on expensive proprietary models or require extensive agentic RL training. To overcome these limitations, we propose Agentic Video Intelligence (AVI), a flexible and training-free framework that can mirror human video comprehension through system-level design and optimization. AVI introduces three key innovations: (1) a human-inspired three-phase reasoning process (Retrieve-Perceive-Review) that ensures both sufficient global exploration and focused local analysis, (2) a structured video knowledge base organized through entity graphs, along with multi-granularity integrated tools, constituting the agent's interaction environment, and (3) an open-source model ensemble combining reasoning LLMs with lightweight base CV models and VLM, eliminating dependence on proprietary APIs or RL training. Experiments on LVBench, VideoMME-Long, LongVideoBench, and Charades-STA demonstrate that AVI achieves competitive performance while offering superior interpretability. |
| title | Agentic Video Intelligence: A Flexible Framework for Advanced Video Exploration and Understanding |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2511.14446 |