Using natural language to write programs is a touchstone problem for computational linguistics. We present an approach that learns to map natural-language descriptions of simple "if-then" rules to executable code. By training and testing on a large corpus of naturally-occurring programs (called "recipes") and their natural language descriptions, we demonstrate the ability to effectively map language to code. We compare a number of semantic parsing approaches on the highly noisy training data collected from ordinary users, and find that loosely synchronous systems perform best.

PDF

In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics (ACL-15), pp. 878--888, Beijing, China, July 2015.

@inproceedings{quirk:acl15, title={Language to Code: Learning Semantic Parsers for If-This-Then-That Recipes}, author={Chris Quirk and Raymond Mooney and Michel Galley}, booktitle={Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics (ACL-15)}, month={July}, address={Beijing, China}, pages={878--888}, url="http://www.cs.utexas.edu/users/ai-labpub-view.php?PubID=127514", year={2015} }

Poster