Proj AutoWriter Paper Reading: Knowledge-based type-aware neural text-to-sql generation

TypeSQL: Knowledge-Based Type-Aware Neural Text-to-SQL Generation

Abstract

  1. 介绍relational db
  2. 介绍Task
    本文: TypeSQL
    方法: 将问题转化为slot filling task
    同时利用了type information
    实验:
    数据集: WikiSQL
    效果: +5.5%的同时减少时间
  3. 显示了读取数据库能极大地增强对不规范查询的效果
  4. 相较于目前的content-sensitive model,+82.6%精度, +17.5%绝对提升
上一篇:[paper reading][EMNLP 2018] Learning Sequence Encoders for Temporal Knowledge Graph Completion


下一篇:Linux 超级补全