Research map / Video / Video-language understanding
TimeSearch-R: Adaptive Temporal Search for Long-Form Video Understanding via Self-Verification Reinforcement Learning
Junwen Pan, Qizhe Zhang, Rui Zhang, Ming Lu, Xin Wan, Yuan Zhang, Chang Liu, Qi She
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 RL post-training with GRPO-CSV after SFT (ablation) | TimeSearch-R · VideoMME (w/o sub) | overall accuracy: 59.2 -> 66.6 | Improved |
| Add SFT cold start to zero-shot CoT search (ablation) | TimeSearch-R · Temporal search | temporal F1: 0.0 -> 7.8 | Improved |
| Add SFT cold start to zero-shot CoT search (ablation) | TimeSearch-R · Temporal search | searched frame completeness: 44.2% -> 60.5% | Improved |
| Add RL after SFT, effect on reasoning consistency (ablation) | TimeSearch-R · Reasoning consistency | consistency rate: +2.6% | Improved |
| Add CSV to GRPO (completeness) (ablation) | GRPO · Temporal search | completeness: 57.2% -> 60.5% | Improved |
| Add CSV to GRPO (temporal F1) (ablation) | GRPO · Temporal search | temporal F1: 7.4 -> 7.8 | Improved |
| Add CSV to GRPO, training stability (ablation) | GRPO · Training dynamics | completeness: training collapses around step 300 without CSV; completeness drops to zero | Improved |
| Add accuracy reward to completeness-only reward (ablation) | GRPO-CSV · QA performance | QA accuracy: completeness reward alone reduces QA accuracy by 0.3% | Improved |
| Apply two-stage data filtering in RL training (Ego+Exo) (ablation) | TimeSearch-R · VideoMME (w/o sub) | overall accuracy: 62.8 -> 66.6 | Improved |
| Add exocentric data to filtered egocentric data (ablation) | TimeSearch-R · VideoMME (w/o sub) | overall accuracy: 65.3 -> 66.6 | Improved |
| Add search without training via CoT prompting to Qwen2.5-VL | Qwen2.5VL-7B · VideoMME (w/o sub) | overall accuracy: 65.1 -> 51.8 | Hurt |
| Train with TimeSearch-R instead of base Qwen2.5-VL | Qwen2.5VL-7B · LongVideoBench val | accuracy: 56.0 -> 60.1 | Improved |
Later papers that compared with this method
2 later accepted papers report TimeSearch-R-7B as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Reflect-R1: Evidence-Driven Reflection for Self-Correction in Long Video Understanding (ECCV 2026 · better on 6 of 6 shared results)
- Searching Videos as Trees: Self-Correcting Agents for Grounded Long Video QA (NeurIPS 2026 · better on 4 of 4 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VideoAgent — VideoAgent: A Memory-augmented Multimodal Agent for Video Understanding (ECCV 2024) · better on 1 of 1 shared result
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
- VideoTrace-R1: Long Video-based Retrieval-Augmented Generation via Reinforcement Learning (ICML 2026 · similarity 0.92)
- Video-R1: Reinforcing Video Reasoning in MLLMs (NeurIPS 2025 · similarity 0.91)
- Find, Fix, Reason: Context Repair for Video Reasoning (ICML 2026 · similarity 0.90)
- Re-thinking Temporal Search for Long-Form Video Understanding (CVPR 2025 · similarity 0.89)
- ReWatch-R1: Boosting Complex Video Reasoning in Large Vision-Language Models through Agentic Data Synthesis (ICLR 2026 · similarity 0.89)
- DeepVideo-R1: Video Reinforcement Fine-Tuning via Difficulty-aware Regressive GRPO (NeurIPS 2025 · similarity 0.88)
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.
