Grape: Knowledge Graph Enhanced Passage Reader for Open-domain Question Answering
Grape: Knowledge Graph Enhanced Passage Reader for Open-domain Question Answering
复制标题
DOI：
10.48550/arxiv.2210.02933
复制
发表时间：
2022-10
期刊：
影响因子：
--
通讯作者：
Mingxuan Ju;W. Yu;Tong Zhao;Chuxu Zhang;Yanfang Ye
中科院分区：
文献类型：
--
作者：
Mingxuan Ju;W. Yu;Tong Zhao;Chuxu Zhang;Yanfang Ye
开放域问答（QA）模型的一个常见思路是采用检索器 - 阅读器流水线，即首先从维基百科中检索出少量相关段落，然后研读这些段落以生成答案。然而，即使是最先进的阅读器也难以捕捉问题与检索到的段落中出现的实体之间的复杂关系，从而导致答案与事实相悖。有鉴于此，我们提出了一种新颖的知识图谱增强型段落阅读器，即Grape，以提升开放域问答中阅读器的性能。具体而言，对于每一对问题与检索到的段落，我们首先构建一个局部二分图，这得益于从阅读器模型中间层提取的实体嵌入。然后，图神经网络在将图和上下文表示融合到阅读器模型的隐藏状态的同时，学习关系知识。在三个开放域问答基准测试上的实验表明，在使用相同的检索器和检索到的段落的情况下，Grape能够将最先进的性能提升高达2.2个精确匹配分数，且开销增加可忽略不计。我们的代码可在https://github.com/jumxglhf/GRAPE上公开获取。
A common thread of open-domain question answering (QA) models employs a retriever-reader pipeline that first retrieves a handful of relevant passages from Wikipedia and then peruses the passages to produce an answer. However, even state-of-the-art readers fail to capture the complex relationships between entities appearing in questions and retrieved passages, leading to answers that contradict the facts. In light of this, we propose a novel knowledge Graph enhanced passage reader, namely Grape, to improve the reader performance for open-domain QA. Specifically, for each pair of question and retrieved passage, we first construct a localized bipartite graph, attributed to entity embeddings extracted from the intermediate layer of the reader model. Then, a graph neural network learns relational knowledge while fusing graph and contextual representations into the hidden states of the reader model. Experiments on three open-domain QA benchmarks show Grape can improve the state-of-the-art performance by up to 2.2 exact match score with a negligible overhead increase, with the same retriever and retrieved passages. Our code is publicly available at https://github.com/jumxglhf/GRAPE.