We introduce a new inference task - Visual Entailment (VE) - which differs
from traditional Textual Entailment (TE) tasks whereby a premise is defined by
an image, rather than a natural language sentence as in TE tasks. A novel
dataset SNLI-VE (publicly available at this https URL) is
proposed for VE tasks based on the Stanford Natural Language Inference corpus
and Flickr30k. We introduce a differentiable architecture called the
Explainable Visual Entailment model (EVE) to tackle the VE problem. EVE and
several other state-of-the-art visual question answering (VQA) based models are
evaluated on the SNLI-VE dataset, facilitating grounded language understanding
and providing insights on how modern VQA based models perform.

本文介绍了一项新的推理任务 - 视觉蕴含（Visual Entailment，VE），VE 与传统的文本蕴含（Textual Entailment，TE）任务不同，它的前提是由图像定义的，而不是像 TE 任务中那样由自然语言句子定义的。在 Stanford 自然语言推理语料库和 Flickr30k 的基础上，提出了一个新的数据集 SNLI-VE，并介绍了一种可解释的视觉蕴含模型（EVE）来解决 VE 问题。此外，本文还将 EVE 和其他几种最先进的基于视觉问答（VQA）的模型在 SNLI-VE 数据集上进行了评估，促进了基于语境的语言理解，并提供了关于现代 VQA 模型性能的见识。