AI Paper Atlas

Research map / Video / Video-language understanding

HowTo100M: Learning a Text-Video Embedding by Watching Hundred Million Narrated Video Clips

Antoine Miech, Dimitri Zhukov, Jean-Baptiste Alayrac, Makarand Tapaswi, Ivan Laptev, Josef Sivic

ICCV 2019 · Video › Video-language understanding · approach: Video-language pretraining · 1,045 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 intra-video negative pairs (ablation)Ours · YouCook2 clip retrievalR@10 (%): 18.1 -> 24.8Improved
add intra-video negative pairs (ablation)Ours · CrossTask step localizationAverage recall (%): 25.7 -> 33.6Improved
add intra-video negative pairs (ablation)Ours · LSMDC clip retrievalR@10 (%): 12.3 -> 14.0Improved
add intra-video negative pairs (ablation)Ours · MSR-VTT clip retrievalR@10 (%): 30.1 -> 29.6Hurt
pre-train on HowTo100M then fine-tune on YouCook2 instead of training on YouCook2 onlyOurs · YouCook2 clip retrievalR@10 (%): 21.5 -> 35.3Improved
train on HowTo100M instead of YouCook2 onlyOurs · YouCook2 clip retrievalR@10 (%): 21.5 -> 24.8Improved
train on HowTo100M instead of MSR-VTT onlyOurs · MSR-VTT clip retrievalR@10 (%): 48.0 -> 29.6Hurt
pre-train on HowTo100M then fine-tune on MSR-VTT instead of training on MSR-VTT onlyOurs · MSR-VTT clip retrievalR@10 (%): 48.0 -> 52.8Improved
pre-train on HowTo100M then fine-tune on LSMDC instead of training on LSMDC onlyOurs · LSMDC clip retrievalR@10 (%): 25.0 -> 27.9Improved
pre-train on HowTo100M then fine-tune on LSMDC instead of training on LSMDC onlyOurs · LSMDC clip retrievalR@1 (%): 7.2 -> 7.1No clear effect
increase training set size of HowTo100M (top 2 to top 200 rank threshold) (ablation)Ours · CrossTask, LSMDC, MSR-VTT, YouCook2average recall / R@10Improved
use only top-scoring positive pairs (max pool rate 0.2 instead of 1.0) (ablation)Ours · MSR-VTT clip retrievalR@10 (%): 29.6 -> 21.9Hurt

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