AI Paper Atlas

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

Alignment Attention by Matching Key and Query Distributions

Shujian Zhang, XINJIE FAN, Huangjie Zheng, Korawat Tanwisuth, Mingyuan Zhou

NeurIPS 2021 · Vision-language and multimodal › Visual question answering · approach: Vision-language model VQA evaluation · 2 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
Replace deterministic attention with Bayesian attention (BAM-WC)GAT · Cora, Citeseer, PubMedaccuracy: 83.00 -> 83.81; 72.50 -> 73.52; 77.26 -> 78.82Improved
Use BAM-WC attention instead of soft attentionMCAN · VQA-v2 validation, noisy imagesaccuracy: 61.25 -> 62.89Improved
Use BAM-WC attention instead of soft attentionMCAN · VQA-v2 validation, original imagesaccuracy: 66.95 -> 67.02No clear effect
Use BAM-WC attention instead of soft attentionMCAN · VQA-v2 validation, original and noisy imagesPAvPU: 70.04 -> 71.21; 65.34 -> 66.75Improved
Use Bayesian attention for image captioningAtt2in · MS-COCOBLEU-4: 32.2 -> 32.8Improved
Use BAM-WC instead of soft attentionLSTM machine translation model · IWSLTBLEU: 32.77 -> 33.81Improved
Fine-tune with BAM-WC attentionALBERT-base · GLUE and SQuADbenchmark scoreMixed
Add a contextual prior rather than use a fixed prior (ablation)BAM on graph classification · Cora, Citeseer, PubMedaccuracy: 83.24 -> 83.34; 72.86 -> 73.04; 78.30 -> 78.76Improved
Use Weibull rather than Lognormal attention distributions (ablation)BAM on graph classification · Cora, Citeseer, PubMedaccuracy: 83.24 -> 83.48; 72.86 -> 73.18; 78.30 -> 78.50Improved
Add a contextual prior rather than use a fixed prior (ablation)BAM on VQA · VQA-v2 validation, original and noisyaccuracy: 66.89 -> 66.93; 61.43 -> 61.58Mixed
Use Weibull rather than Lognormal attention distributions (ablation)BAM on VQA · VQA-v2 validation, original and noisyaccuracy: 66.89 -> 66.93; 61.43 -> 61.60Mixed
Use semi-analytic KL regularization rather than fully sampled KL (ablation)BAM · MS-COCO image captioningtraining stabilityImproved

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