AI Paper Atlas

Research map / Video / Video-language understanding

Prompting Visual-Language Models for Efficient Video Understanding

Chen Ju, Tengda Han, Kunhao Zheng, Ya Zhang, Weidi Xie

ECCV 2022 · Video › Video-language understanding · approach: Video-language pretraining · 458 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
increase learnable prompt vectors from 2+X+2 to 16+X+16 (ablation)CLIP prompt learning · Kinetics-400 closed-setTOP1 accuracy: 65.4 -> 68.8Improved
add one temporal Transformer layer on frame features (ablation)CLIP prompt learning (16+X+16) · Kinetics-400 closed-setTOP1 accuracy: 68.8 -> 75.8Improved
go from 2 to 4 temporal Transformer layers (ablation)CLIP prompt learning (16+X+16) · Kinetics-700 closed-setaverage of TOP1 and TOP5: 76.6 -> 76.4No clear effect
learn prompt vectors instead of handcrafted template (ablation)CLIP · Kinetics-700 closed-setaverage of TOP1 and TOP5: 52.4 -> 70.5Improved
replace linear probe with learned prompts (16K parameters) (ablation)CLIP image encoder · Kinetics-700 closed-setaverage of TOP1 and TOP5: 66.1 -> 70.5Improved
add 4-layer temporal Transformer to prompted CLIP (ablation)CLIP prompt learning · Something-Something V2 closed-setTOP1 accuracy: 18.4 -> 38.1Improved
add temporal Transformer in 5-shot 5-way few-shot recognition (ablation)CLIP prompt learning · UCF-101 few-shot 5-wayTOP1 accuracy: 98.3 -> 97.8No clear effect
add temporal Transformer in 5-shot all-class few-shot recognition (ablation)CLIP prompt learning · UCF-101 few-shot 5-shot, all classesTOP1 accuracy: 77.6 -> 79.5Improved
learn prompts instead of handcrafted template in few-shot (ablation)CLIP · UCF-101 few-shot 5-shot, all classesTOP1 accuracy: 64.7 -> 77.6Improved
add temporal Transformer layers for zero-shot recognition (ablation)CLIP prompt learning (16+X+16) · Kinetics-700 zero-shotTOP1 accuracy: 58.4 -> 45.5Hurt
increase prompt vectors from 4+X+4 to 32+X+32 in zero-shot recognition (ablation)CLIP prompt learning · Kinetics-700 zero-shotTOP1 accuracy: 57.4 -> 57.5No clear effect
learn prompts instead of handcrafted template for zero-shot recognition (ablation)CLIP · Kinetics-700 zero-shotTOP1 accuracy: 52.4 -> 58.4Improved

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