AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

Learning Models for Actions and Person-Object Interactions with Transfer to Question Answering

Arun Mallya, Svetlana Lazebnik

ECCV 2016 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA · 118 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 weighted loss (ablation)Fusion-1 · HICO testmAP: 33.6 -> 36.0Improved
add weighted loss (ablation)Fusion-2 · HICO testmAP: 33.8 -> 36.1Improved
fuse person box and full image features early (Fusion-2) instead of full image onlyVGG-16 · HICO testmAP: 29.4 -> 33.8Improved
fuse features early (Fusion-2) instead of averaging independent predictions (Scene-RCNN)VGG-16 · HICO testmAP: 29.0 -> 33.8Improved
use person bounding box only instead of adding full image (ablation)VGG-16 with MIL · HICO testmAP: 29.4 -> 14.6Hurt
Fusion-2 channel reduction before stacking versus Fusion-1Fusion · HICO testmAP: 33.6 -> 33.8Improved
replace R*CNN secondary box search with full image box (Scene-RCNN)VGG-16 · HICO testmAP: 28.5 -> 29.0Improved
swap AlexNet for VGG-16 backbone with full imagefull-image classifier · HICO testmAP: 19.4 -> 29.4Improved
train with MIL over ground truth instances instead of a label per instance (ablation)Fusion-2 · MPII testmAP: 32.24 -> 31.89Hurt
use detected instead of ground truth person boxes (ablation)Fusion-2 · MPII testmAP: 32.24 -> 32.02No clear effect
replace ImageNet VGG-16 fc7 with HICO full-image network fc7MadLibs CCA · MadLibs Easy Person's Activityaccuracy: 80.79 -> 86.54Improved
use Fusion-2 on person boxes instead of HICO full-image fc7MadLibs CCA · MadLibs Person's Activity (Filtered Hard)accuracy: 72.39 -> 73.46Improved

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