AI Paper Atlas

Research map / Video / Video-language understanding

Multi-Stage Aggregated Transformer Network for Temporal Language Localization in Videos

Mingxing Zhang, Yang Yang, Xinghan Chen, Yanli Ji, Xing Xu, Jingjing Li, Heng Tao Shen

CVPR 2021 · Video › Video-language understanding · approach: Dense video captioning · 86 citations (OpenAlex)

Check 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
Use a visual-language transformer rather than Conv-LSTM with temporal convolution (ablation)Temporal language localization model · ActivityNet CaptionsRank 1 @ IoU=0.5: 41.12 -> 46.96Improved
Add multi-stage aggregation rather than convolutional candidate features (ablation)Conv-LSTM model · ActivityNet CaptionsRank 1 @ IoU=0.5: 41.12 -> 45.16Improved
Add multi-stage aggregation rather than mean pooling (ablation)Visual-language transformer · ActivityNet CaptionsRank 1 @ IoU=0.5: 45.92 -> 46.96Improved
Add multi-stage aggregation rather than stacked convolution (ablation)Visual-language transformer · ActivityNet CaptionsRank 1 @ IoU=0.5: 45.36 -> 46.96Improved
Add multi-stage aggregation rather than RoI pooling (ablation)Visual-language transformer · ActivityNet CaptionsRank 1 @ IoU=0.5: 45.02 -> 46.96Improved
Decouple transformer parameters for visual and language modalities (ablation)Visual-language transformer with MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.96 -> 48.02Improved
Use masked word prediction during training (ablation)De-VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.26 -> 48.02Improved
Use three stages instead of only starting and ending stages (ablation)De-VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.84 -> 48.02Improved
Use three stages rather than four moment stages (ablation)De-VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 47.69 -> 48.02Improved
Increase shared-transformer parameters 4 times in width (ablation)VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.96 -> 45.52Hurt
Increase shared-transformer parameters 4 times in depth (ablation)VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.96 -> 20.74Hurt
Use two BERT streams instead of a decoupled single transformer (ablation)De-VLTrans-MSA · ActivityNet CaptionsRank 1 @ IoU=0.5: 46.74 -> 48.02Hurt

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