视觉和语言的推理：探索补充知识的好处
Reasoning over Vision and Language: Exploring the Benefits of Supplemental Knowledge
摘要：
视觉和语言模型的适用性限制由其训练数据的覆盖范围定义。像视觉问答（VQA）这样的任务通常需要常识和事实信息，而不是从特定于任务的数据集中学到的信息。本文研究了从通用知识库（KBs）到视觉和语言Transformers的知识注入。我们使用一个辅助训练目标，鼓励学习的表示与知识库中匹配实体的图嵌入对齐。我们实证研究的相关性，各种知识库多个任务和基准。该技术带来了明显的好处，知识需求的问答任务（OK-VQA，FVQA），通过捕捉语义和关系知识缺乏现有的模型。更令人惊讶的是，该技术也有利于视觉推理任务（NLVR 2，SNLI-VE）。我们进行探测实验，并表明，注入额外的知识规则化的嵌入空间，提高了词汇和语义相似性的表示。该技术与模型无关，可以以最小的计算开销扩展任何视觉和语言Transformer的适用性。
英文摘要：
The limits of applicability of vision-and-language models are defined by the coverage of their training data. Tasks like vision question answering (VQA) often require commonsense and factual information beyond what can be learned from task-specific datasets. This paper investigates the injection of knowledge from general-purpose knowledge bases (KBs) into vision-and-language transformers. We use an auxiliary training objective that encourages the learned representations to align with graph embeddings of matching entities in a KB. We empirically study the relevance of various KBs to multiple tasks and benchmarks. The technique brings clear benefits to knowledge-demanding question answering tasks (OK-VQA, FVQA) by capturing semantic and relational knowledge absent from existing models. More surprisingly, the technique also benefits visual reasoning tasks (NLVR2, SNLI-VE). We perform probing experiments and show that the injection of additional knowledge regularizes the space of embeddings, which improves the representation of lexical and semantic similarities. The technique is model-agnostic and can expand the applicability of any vision-and-language transformer with minimal computational overhead.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
