AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Look at What I’m Doing: Self-Supervised Spatial Grounding of Narrations in Instructional Videos

Reuben Tan, Bryan A. Plummer, Kate Saenko, Hailin Jin, Bryan Russell

NeurIPS 2021 · Vision-language and multimodal › Cross-modal learning · approach: Unimodal-to-multimodal representation transfer · 13 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
Use sentence-level rather than word-level contrastive loss (ablation)CoMMA · YouCook2-Interactionslocalization accuracy (%): 34.91 -> 55.80Improved
Use CoMMA sentence-level grounding instead of the strongest listed baseline, Shallow AttentionYouCook2-Interactions · YouCook2-Interactionslocalization accuracy (%): 48.30 -> 55.80Improved
Self-train CoMMA on narrated video clips for object localizationCoMMA · YouCook2-BB validationfull localization (%): 46.95 -> 59.25Improved
Add a second cross-attention layer (ablation)CoMMA · YouCook2-Interactionslocalization accuracy (%): 48.30 -> 51.09Improved
Add full spatiotemporal self-attention instead of spatial-only self-attention (ablation)CoMMA · YouCook2-Interactionslocalization accuracy (%): 53.88 -> 55.80Improved
Use temporal-only self-attention instead of full spatiotemporal self-attention (ablation)CoMMA · YouCook2-Interactionslocalization accuracy (%): 52.72 -> 55.80Hurt
Use separate spatial and temporal self-attention instead of full spatiotemporal self-attention (ablation)CoMMA · YouCook2-Interactionslocalization accuracy (%): 52.15 -> 55.80Hurt
Add a second cross-attention layer to sentence-loss-only CoMMA (ablation)CoMMA hybrid phrase-grounding model · Flickr30KR@1: 35.43 -> 39.03Improved
Add self-attention to sentence-loss-only CoMMA (ablation)CoMMA hybrid phrase-grounding model · Flickr30KR@1: 39.03 -> 39.31Improved
Add sentence-level loss to the one-layer word-loss hybrid (ablation)CoMMA hybrid phrase-grounding model · Flickr30KR@1: 49.99 -> 52.43Improved
Add a second cross-attention layer to the word + sentence loss hybrid (ablation)CoMMA hybrid phrase-grounding model · Flickr30KR@1: 52.43 -> 53.80Improved
Add self-attention to the two-layer word + sentence loss hybrid (ablation)CoMMA hybrid phrase-grounding model · Flickr30KR@1: 53.80 -> 48.53Hurt

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