AI Paper Atlas

Research map / Video / Video-language understanding

VideoTemp-o3: Harmonizing Temporal Grounding and Video Understanding in Agentic Thinking-with-Videos

Wenqi Liu, Yunxiao Wang, Shijie Ma, Meng Liu, Qile Su, Tianke Zhang, Haonan fan, Changyi Liu, Kaiyu Jiang, Jiankang Chen, Kaiyu Tang, Bin Wen and 5 more

ICML 2026 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
apply RL after SFTVideoTemp-o3-7B · VideoMME Overallaccuracy: 60.6 -> 64.5Improved
add grounding data to SFT (ablation)VideoTemp-o3 · ReXTimemIoU: 13.0 -> 29.5Improved
add grounding data to SFT (ablation)VideoTemp-o3 · VideoMME Overallaccuracy: 63.0 -> 64.5Improved
add unified masking (supervise only last two turns) (ablation)VideoTemp-o3 SFT · VideoMMMU Overallaccuracy: 47.9 -> 53.2Improved
add unified masking (ablation)VideoTemp-o3 SFT · ReXTimemIoU: 18.8 -> 29.5Improved
add IoU reward in RL (ablation)VideoTemp-o3 RL · VideoMMMU Overallaccuracy: 51.6 -> 53.2Improved
use penalty-aware IoU reward instead of plain IoU reward (ablation)VideoTemp-o3 RL · VideoMMMU Overallaccuracy: 44.2 -> 53.2Improved
use penalty-aware IoU reward instead of plain IoU reward (ablation)VideoTemp-o3 RL · ReXTimemIoU: 23.8 -> 29.5Improved
learned on-demand cropping instead of always cropping once (ablation)VideoTemp-o3 · VideoMME (w/o subtitle) Overallaccuracy: 63.0 -> 64.5Improved
learned tool use instead of answering directly without tool call (ablation)VideoTemp-o3 · VideoMME (w/o subtitle) Overallaccuracy: 62.1 -> 64.5Improved
add prompt-only re-grounding to the base modelQwen2.5-VL-7B · VideoMME (w/o subtitle) Overallaccuracy: 59.9 -> 58.4No clear effect
add inference-time self-correction to the base modelQwen2.5-VL-7B · VideoMME (w/o subtitle) Overallaccuracy: 59.9 -> 60.5No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy