VQA-E stands for Visual Question Answering with Explanation. It is a visual question answering dataset with explanations. The model involved needs to predict and generate answer explanations. VQA v2 datasetAutomatically derived, a textual analysis is synthesized for each “image-question-answer triad”, making the question-answering process easier to understand and trace.