AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Does Visual Pretraining Help End-to-End Reasoning?

Chen Sun, Calvin Luo, Xingyi Zhou, Anurag Arnab, Cordelia Schmid

NeurIPS 2023 · Detection and segmentation › Open-set and few-shot detection · approach: Visual relationship and scene graph detection · 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
Replace supervised visual pretraining with IV-CL self-supervised video pretrainingViT-B visual encoder · CATERaccuracy (%): 56.64% -> 70.14 (±0.59)%Improved
Replace supervised visual pretraining with IV-CL self-supervised video pretrainingViT-B visual encoder · ACRE compositionalityaccuracy (%): 67.27% -> 93.27 (±0.22)%Improved
Use object detection rather than classification pretrainingViT-B visual encoder · CATERaccuracy (%): 41.48% -> 56.64%Improved
Use larger-scale JFT classification pretraining instead of ImageNet-21k classification pretrainingViT-B visual encoder · ACRE compositionalityaccuracy (%): 60.73% -> 48.32%Hurt
Use IV-CL instead of video MAE pretrainingEnd-to-end visual reasoning model · CATER staticTop-1 accuracy (%): 63.7 -> 70.1 ± 0.6Improved
Use IV-CL instead of image MAE pretrainingEnd-to-end visual reasoning model · ACRE compositionalityaccuracy (%): 80.27 -> 93.27 ± 0.22Improved
Use 37.5% rather than 12.5% masking (ablation)IV-CL · CATERaccuracy (%): 66.35% -> 70.14%Improved
Use 37.5% rather than 50% masking (ablation)IV-CL · CATERaccuracy (%): 66.57% -> 70.14%Improved
Use 37.5% rather than 87.5% masking (ablation)IV-CL · CATERaccuracy (%): 61.94% -> 70.14%Improved
Add context frames to reconstruction rather than using none (ablation)IV-CL · CATERaccuracy (%): 65.35% -> 70.14%Improved
Use 8 rather than 4 context frames (ablation)IV-CL · CATERaccuracy (%): 67.47% -> 70.14%Improved
Use 8 rather than 16 context frames (ablation)IV-CL · CATERaccuracy (%): 64.34% -> 70.14%Improved

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