AI Paper Atlas

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

Free VQA Models from Knowledge Inertia by Pairwise Inconformity Learning

Yiyi Zhou, Rongrong Ji, Jinsong Su, Xiangming Li, Xiaoshuai Sun

AAAI 2019 · Vision-language and multimodal › Visual question answering · approach: Attention and prior-robust VQA · 5 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 answer embedding loss only (ablation)Baseline · VQA2.0 valaccuracy (All): 60.37 -> 60.74Improved
add full Pairwise Inconformity Learning (ablation)Baseline · VQA2.0 valaccuracy (All): 60.37 -> 61.56Improved
add hard example mining to PIL (ablation)Baseline+PIL · VQA2.0 valaccuracy (All): 61.56 -> 62.01Improved
add PIL with Faster R-CNN features (ablation)Baseline+FRCNN · VQA2.0 valaccuracy (All): 63.72 -> 64.58Improved
add hard example mining with FRCNN features (ablation)Baseline+PIL+FRCNN · VQA2.0 valaccuracy (All): 64.58 -> 64.63No clear effect
increase beta from 0.25 to 1 (alpha=0.01) (ablation)Baseline+PIL · VQA2.0 valaccuracy (All): 61.56 -> 61.06Hurt
decrease beta to 0.05 (alpha=0.01) (ablation)Baseline+PIL · VQA2.0 valaccuracy (All): 61.56 -> 60.84Hurt
set alpha=0.25 (beta=0.25) (ablation)Baseline+PIL · VQA2.0 valaccuracy (All): 61.56 -> 61.50No clear effect
l2 regularization on joint features instead of PIL (ablation)Baseline · VQA2.0 valaccuracy (All): 60.37 -> 59.47Hurt
contrastive loss on joint features (ablation)Baseline · VQA2.0 valaccuracy (All): 60.37 -> 59.58Hurt
KL-divergence regularization on predictions (ablation)Baseline · VQA2.0 valaccuracy (All): 60.37 -> 59.18Hurt
apply PIL on attention feature fv instead of joint feature (ablation)Baseline+PIL · VQA2.0 valaccuracy (All): 61.56 -> 61.32Hurt

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