Table Search Using a Deep Contextualized Language Model
Table Search Using a Deep Contextualized Language Model
复制标题
DOI：
10.1145/3397271.3401044
复制
发表时间：
2020-05
期刊：
影响因子：
--
通讯作者：
Zhiyu Chen;M. Trabelsi;J. Heflin;Yinan Xu;Brian D. Davison
中科院分区：
文献类型：
--
作者：
Zhiyu Chen;M. Trabelsi;J. Heflin;Yinan Xu;Brian D. Davison
BERT等预先训练的语境化语言模型在各种自然语言处理基准上取得了令人印象深刻的结果。得益于多个预训练任务和大规模的训练语料库，预训练模型可以捕获复杂的句法词汇关系。在本文中，我们使用深层上下文语言模型BERT来执行即席表格检索任务。考虑到ERT的表结构和输入长度限制，研究了如何对表内容进行编码。我们还提出了一种方法，该方法结合了以前关于表格检索的文献中的特征，并与BERT联合训练它们。在公开数据集上的实验表明，在不同的评价指标下，我们最好的方法可以在很大程度上超过现有的方法和BERT基线。
Pretrained contextualized language models such as BERT have achieved impressive results on various natural language processing benchmarks. Benefiting from multiple pretraining tasks and large scale training corpora, pretrained models can capture complex syntactic word relations. In this paper, we use the deep contextualized language model BERT for the task of ad hoc table retrieval. We investigate how to encode table content considering the table structure and input length limit of BERT. We also propose an approach that incorporates features from prior literature on table retrieval and jointly trains them with BERT. In experiments on public datasets, we show that our best approach can outperform the previous state-of-the-art method and BERT baselines with a large margin under different evaluation metrics.