AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

InstAP: Instance-Aware Vision-Language Pre-Train for Spatial-Temporal Understanding

Ashutosh Kumar, Rajat Saini, Jingjing Pan, Mustafa Erdogan, Mingfang Zhang, Betty Le Dem, Norimasa Kobori, Quan Kong

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 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
add instance-aware loss L_inst (ablation)InstAP · InstVL-1K (video) instancemean recall: 57.71 -> 75.32Improved
add instance-aware loss L_inst (ablation)InstAP · InstVL-1K (img-zero) instancemean recall: 49.98 -> 63.94Improved
add instance-aware loss L_inst (ablation)InstAP · InstVL-1K (video) globalmean recall: 91.55 -> 97.03Improved
add instance-aware loss L_inst (ablation)InstAP · DiDeMomean recall: 65.98 -> 70.01Improved
add learnable instance temperature (ablation)InstAP · InstVL-1K (img)mean recall: 59.10 -> 67.19Improved
add weighted instance loss (lambda 0.1) (ablation)InstAP · InstVL-1K (video)mean recall: 55.22 -> 58.16Improved
add caption sub-sampling (ablation)InstAP · InstVL-1K (img)mean recall: 68.17 -> 71.65Improved
add 50K instance video trajectory data (ablation)InstAP · InstVL-1K (video)mean recall: 58.97 -> 75.32Improved
treat instance captions as global captions (g+i) vs instance-aware trainingUMT-L baseline · InstVL-10K (img)T2V R@1: 44.05 -> 34.83Hurt
fine-tune UMT-L on InstVL global captions onlyUMT-L · MSR-VTTR@1: 39.7 -> 35.4Hurt
use InstAP instead of UMT-Lzero-shot retrieval · DiDeMoR@1: 47.0 -> 54.0Improved
use InstAP instead of UMT-Lzero-shot retrieval · LSMDCR@1: 26.0 -> 23.5Hurt

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