AI Paper Atlas

Research map / Video / Video-language understanding

Learning to Refuse: Refusal-Aware Reinforcement Fine-Tuning for Hard-Irrelevant Queries in Video Temporal Grounding

Jin-Seop Lee, SungJoon Lee, SeongJun Jung, Boyang Li, Jee-Hyong Lee

CVPR 2026 · Video › Video-language understanding · approach: Temporal video grounding · 2 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
train GRPO on simply-shuffled queries only (no HI-VTG) (ablation)Time-R1 · HI-ActivityNet (Time-R1)F1 avg: 70.5 -> 70.0No clear effect
add HI-VTG hard-irrelevant data with refuse-IoU reward (ablation)Time-R1 · HI-ActivityNet (Time-R1)F1 avg: 70.0 -> 75.1Improved
add explain reward (ablation)Time-R1 + GRPO R.IoU · HI-ActivityNet (Time-R1)F1 avg: 75.1 -> 75.6Improved
add explain reward (ablation)Time-R1 + GRPO R.IoU · HI-ActivityNet (Time-R1)LLM explanation score: 2.29 -> 2.37Improved
add query correction reward (ablation)Time-R1 + GRPO R.IoU-Exp · HI-ActivityNet (Time-R1)F1 avg: 75.6 -> 76.3Improved
add query correction reward (ablation)Time-R1 + GRPO R.IoU-Exp · HI-ActivityNet (Time-R1)RA-IoU mIoU: 52.2 -> 51.9No clear effect
replace RA-RFT with SFT on HI-VTG (ablation)Time-R1 · HI-ActivityNet (Time-R1)RA-IoU mIoU: 51.9 -> 20.9Hurt
add RA-RFTTime-R1 · HI-ActivityNetRA-IoU mIoU: 45.9 -> 51.9Improved
add RA-RFTVideoChat-R1 · HI-ActivityNetRA-IoU mIoU: 39.4 -> 46.5Improved
add RA-RFTTime-R1 · HI-TVGBenchRA-IoU mIoU: 37.3 -> 46.6Improved
add RA-RFTQwen2.5-VL-7B · HI-ActivityNetRA-IoU mIoU: 27.8 -> 49.6Improved
add RA-RFTTime-R1 · HI-ActivityNet (strong hard-irrelevant)F1: 60.0 -> 70.6Improved

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.

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

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