Papers › Tips and Tricks for Visual Question Answering: Learnings from the 2017 Challenge

Tips and Tricks for Visual Question Answering: Learnings from the 2017 Challenge

9 Aug 2017CVPR 2018 6arXiv:1708.02711archive 2025-07-28

Damien Teney, Peter Anderson, Xiaodong He, Anton Van Den Hengel

This paper presents a state-of-the-art model for visual question answering (VQA), which won the first place in the 2017 VQA Challenge. VQA is a task of significant importance for research in artificial intelligence, given its multimodal nature, clear evaluation protocol, and potential real-world applications. The performance of deep neural networks for VQA is very dependent on choices of architectures and hyperparameters. To help further research in the area, we describe in detail our high-performing, though relatively simple model. Through a massive exploration of architectures and hyperparameters representing more than 3,000 GPU-hours, we identified tips and tricks that lead to its success, namely: sigmoid outputs, soft training targets, image features from bottom-up attention, gated tanh activations, output embeddings initialized using GloVe and Google Images, large mini-batches, and smart shuffling of training data. We provide a detailed analysis of their impact on performance to assist others in making an appropriate selection.

PaperPDFConference PDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

feifengwhu/question_attention mentioned on GitHubpytorch report
hengyuan-hu/bottom-up-attention-vqa mentioned on GitHubpytorch report
shailzajolly/icdar_vqa mentioned on GitHubpytorch report
snagiri/ECE285_Jarvis_ProjectA mentioned on GitHubpytorch report
thilinicooray/Bottom-up-vqa mentioned on GitHubpytorch report
yangdsh/VQA-BUTD-demo mentioned on GitHub report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Visual Question AnsweringVisual Question Answering (VQA)

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Visual Question Answering (VQA) VQA v2 test-dev Image features from bottom-up attention (adaptive K, ensemble) Accuracy 69.87 #33 of 56 Archive leaderboard report
Visual Question Answering (VQA) VQA v2 test-std Image features from bottom-up attention (adaptive K, ensemble) overall 70.3 #29 of 38 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

GloVe

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections