数据库语言：学习数据库自然语言接口的弱监督
DBPal: Weak Supervision for Learning a Natural Language Interface to Databases
摘要：
本文介绍了一种新的系统DBSTOM，它使用神经机器翻译模型将自然语言语句翻译成SQL语句。虽然其他最近的方法使用神经机器翻译来实现数据库的自然语言接口（NLIDB），但是现有技术依赖于具有手动策划的训练数据的监督学习，这导致用于支持每个新数据库模式的大量开销。为了避免这个问题，DBSQL实现了一种基于弱监督的新型训练管道，该管道综合了来自给定数据库模式的所有训练数据。在我们的评估中，我们表明DBDB可以优于现有的基于规则的NLIDB，同时实现与其他NLIDB相当的性能，这些NLIDB利用深度神经网络模型，而无需依赖于手动策划的训练数据。
英文摘要：
This paper describes DBPal, a new system to translate natural language utterances into SQL statements using a neural machine translation model. While other recent approaches use neural machine translation to implement a Natural Language Interface to Databases (NLIDB), existing techniques rely on supervised learning with manually curated training data, which results in substantial overhead for supporting each new database schema. In order to avoid this issue, DBPal implements a novel training pipeline based on weak supervision that synthesizes all training data from a given database schema. In our evaluation, we show that DBPal can outperform existing rule-based NLIDBs while achieving comparable performance to other NLIDBs that leverage deep neural network models without relying on manually curated training data for every new database schema.
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。
