Research map / Video / Video-language understanding
VCA: Video Curious Agent for Long Video Understanding
Zeyuan Yang, Delin Chen, Xueyang Yu, Maohao Shen, Chuang Gan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add intrinsic reward model to guide exploration (ablation) | VCA · LVBench | accuracy: 39.5 -> 41.3 | Improved |
| add intrinsic reward model to guide exploration (ablation) | VCA · EgoSchema subset | accuracy: 72.6 -> 73.6 | Improved |
| add tree-search segment exploration (ablation) | VCA · LVBench | accuracy: 36.2 -> 41.3 | Improved |
| add tree-search segment exploration (ablation) | VCA · EgoSchema subset | accuracy: 69.8 -> 73.6 | Improved |
| increase memory buffer from 8 to 32 frames (ablation) | VCA · LVBench | accuracy: 38.9 -> 43.2 | Improved |
| use segment-semantic reward instead of image relevance (ablation) | VCA · EgoSchema subset (Qwen2-VL) | accuracy: 74.0 -> 75.2 | Improved |
| use ground-truth reward scores (ablation) | VCA · LVBench | accuracy: 41.3 -> 44.2 | Improved |
| replace uniform sampling with VCA exploration | GPT-4o · LVBench | accuracy: 34.7 -> 41.3 | Improved |
| replace dense-caption agents with VCA | VideoAgent/VideoTree · EgoSchema subset | accuracy: 67.0 -> 73.6 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- VideoTree — VideoTree: Adaptive Tree-based Video Representation for LLM Reasoning on Long Videos (CVPR 2025) · better on 3 of 3 shared results
- VideoAgent — VideoAgent: A Memory-augmented Multimodal Agent for Video Understanding (ECCV 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Deep Video Discovery: Agentic Search with Tool Use for Long-form Video Understanding (NeurIPS 2025 · similarity 0.89)
- VideoSeek: Long-Horizon Video Agent with Tool-Guided Seeking (CVPR 2026 · similarity 0.88)
- Think, Then Verify: A Hypothesis-Verification Multi-Agent Framework for Long Video Understanding (CVPR 2026 · similarity 0.88)
- A Multi-Agent Perception-Action Alliance for Efficient Long Video Reasoning (CVPR 2026 · similarity 0.87)
- ReAgent-V: A Reward-Driven Multi-Agent Framework for Video Understanding (NeurIPS 2025 · similarity 0.87)
- Searching Videos as Trees: Self-Correcting Agents for Grounded Long Video QA (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
