Cross-domain Semantic Parsing via Paraphrasing

Abstract

Existing studies on semantic parsing mainly focus on the in-domain setting. We formulate cross-domain semantic parsing as a domain adaptation problem: train a semantic parser on some source domains and then adapt it to the target domain. Due to the diversity of logical forms in different domains, this problem presents unique and intriguing challenges. By converting logical forms into canonical utterances in natural language, we reduce semantic parsing to paraphrasing, and develop an attentive sequence-to-sequence paraphrase model that is general and flexible to adapt to different domains. We discover two problems, small micro variance and large macro variance, of pre-trained word embeddings that hinder their direct use in neural networks, and propose standardization techniques as a remedy. On the popular Overnight dataset, which contains eight domains, we show that both cross-domain training and standardized pre-trained word embeddings can bring significant improvement.

12 pages, 2 figures, accepted by EMNLP2017

References in corpus (8)

• Sequence to Sequence Learning with Neural Networks
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Improving neural networks by preventing co-adaptation of feature detectors
• Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
• Frustratingly Easy Domain Adaptation
• Convolutional Neural Network Architectures for Matching Natural Language Sentences
• Language to Logical Form with Neural Attention
• Universal Semantic Parsing

Cited by in corpus (10)

• SyntaxSQLNet: Syntax Tree Networks for Complex and Cross-DomainText-to-SQL Task
• Exploiting Rich Syntactic Information for Semantic Parsing with Graph-to-Sequence Model
• Paraphrase Generation with Deep Reinforcement Learning
• A Transfer-Learnable Natural Language Interface for Databases
• ParaSCI: A Large Scientific Paraphrase Dataset for Longer Paraphrase Generation
• Zero-shot Transfer Learning for Semantic Parsing
• Transferable Natural Language Interface to Structured Queries aided by Adversarial Generation
• Domain Adaptation for Semantic Parsing
• Few-Shot Semantic Parsing for New Predicates
• Knowledge Graph Simple Question Answering for Unseen Domains