# Rethinking Data Augmentation for Robust Visual Question Answering

Source: [Rethinking Data Augmentation for Robust Visual Question Answering](<https://www.sota2.com/research/paper/68fa6d681e209131cea39954>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T06:17:30.291Z

Authors: Long Chen, Yuhang Zheng, Jun Xiao

Published: 2022-07-18T16:30:50.000Z

## Abstract

Data Augmentation (DA) -- generating extra training samples beyond original training set -- has been widely-used in today's unbiased VQA models to mitigate the language biases. Current mainstream DA strategies are synthetic-based methods, which synthesize new samples by either editing some visual regions/words, or re-generating them from scratch. However, these synthetic samples are always unnatural and error-prone. To avoid this issue, a recent DA work composes new augmented samples by randomly pairing pristine images and other human-written questions. Unfortunately, to guarantee augmented samples have reasonable ground-truth answers, they manually design a set of heuristic rules for several question types, which extremely limits its generalization abilities. To this end, we propose a new Knowledge Distillation based Data Augmentation for VQA, dubbed KDDAug. Specifically, we first relax the requirements of reasonable image-question pairs, which can be easily applied to any question types. Then, we design a knowledge distillation (KD) based answer assignment to generate pseudo answers for all composed image-question pairs, which are robust to both in-domain and out-of-distribution settings. Since KDDAug is a model-agnostic DA strategy, it can be seamlessly incorporated into any VQA architectures. Extensive ablation studies on multiple backbones and benchmarks have demonstrated the effectiveness and generalization abilities of KDDAug.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Visual Question Answering on VQA-CP v2 (test)](<https://www.sota2.com/research/sota/visual-question-answering-on-vqa-cp-v2-test.md>): Visual Question Answering; VQA-CP v2 (test); Overall Accuracy: 60.24.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2207.08739>)

[PDF](<https://arxiv.org/pdf/2207.08739.pdf>)