Counterfactual VQA: A Cause-Effect Look at Language Bias

Understand

VQA models may tend to rely on language bias as a shortcut and thus fail to sufficiently learn the multi-modal knowledge from both vision and language.

• Recent debiasing methods proposed to exclude the language prior during inference.
• However, they fail to disentangle the "good" language context and "bad" language bias from the whole.
• In this paper, we investigate how to mitigate language bias in VQA.