A Survey on Semantic Parsing
Aishwarya Kamath, Rajarshi Das
TL;DR
This survey maps the evolution of semantic parsing from rule-based to neural approaches, focusing on how natural language is converted into executable meaning representations across diverse contexts. It catalogs representation formalisms (logic, graphs, and programming languages), grammar designs, and supervision strategies, including denotation-based and weak supervision, as well as end-to-end Seq2Seq models with structure-aware decoding. The article also covers learning paradigms (MML, RL, MMR), intermediate representations, and code-generation ventures, highlighting methods to manage search, data scarcity, and cross-domain generalization. By connecting historical milestones with contemporary trends, it provides a framework for building robust, scalable semantic parsers applicable to KB querying, robotics, and intelligent assistants. The discussed future directions emphasize confidence, domain transfer, and semantically grounded evaluation to advance practical deployments.
Abstract
A significant amount of information in today's world is stored in structured and semi-structured knowledge bases. Efficient and simple methods to query them are essential and must not be restricted to only those who have expertise in formal query languages. The field of semantic parsing deals with converting natural language utterances to logical forms that can be easily executed on a knowledge base. In this survey, we examine the various components of a semantic parsing system and discuss prominent work ranging from the initial rule based methods to the current neural approaches to program synthesis. We also discuss methods that operate using varying levels of supervision and highlight the key challenges involved in the learning of such systems.