Learning to Infer Program Sketches

Maxwell Nye, Luke Hewitt, Joshua Tenenbaum, Armando Solar-Lezama
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:4861-4870, 2019.

Abstract

Our goal is to build systems which write code automatically from the kinds of specifications humans can most easily provide, such as examples and natural language instruction. The key idea of this work is that a flexible combination of pattern recognition and explicit reasoning can be used to solve these complex programming problems. We propose a method for dynamically integrating these types of information. Our novel intermediate representation and training algorithm allow a program synthesis system to learn, without direct supervision, when to rely on pattern recognition and when to perform symbolic search. Our model matches the memorization and generalization performance of neural synthesis and symbolic search, respectively, and achieves state-of-the-art performance on a dataset of simple English description-to-code programming problems.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-nye19a, title = {Learning to Infer Program Sketches}, author = {Nye, Maxwell and Hewitt, Luke and Tenenbaum, Joshua and Solar-Lezama, Armando}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {4861--4870}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/nye19a/nye19a.pdf}, url = {https://proceedings.mlr.press/v97/nye19a.html}, abstract = {Our goal is to build systems which write code automatically from the kinds of specifications humans can most easily provide, such as examples and natural language instruction. The key idea of this work is that a flexible combination of pattern recognition and explicit reasoning can be used to solve these complex programming problems. We propose a method for dynamically integrating these types of information. Our novel intermediate representation and training algorithm allow a program synthesis system to learn, without direct supervision, when to rely on pattern recognition and when to perform symbolic search. Our model matches the memorization and generalization performance of neural synthesis and symbolic search, respectively, and achieves state-of-the-art performance on a dataset of simple English description-to-code programming problems.} }
Endnote
%0 Conference Paper %T Learning to Infer Program Sketches %A Maxwell Nye %A Luke Hewitt %A Joshua Tenenbaum %A Armando Solar-Lezama %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-nye19a %I PMLR %P 4861--4870 %U https://proceedings.mlr.press/v97/nye19a.html %V 97 %X Our goal is to build systems which write code automatically from the kinds of specifications humans can most easily provide, such as examples and natural language instruction. The key idea of this work is that a flexible combination of pattern recognition and explicit reasoning can be used to solve these complex programming problems. We propose a method for dynamically integrating these types of information. Our novel intermediate representation and training algorithm allow a program synthesis system to learn, without direct supervision, when to rely on pattern recognition and when to perform symbolic search. Our model matches the memorization and generalization performance of neural synthesis and symbolic search, respectively, and achieves state-of-the-art performance on a dataset of simple English description-to-code programming problems.
APA
Nye, M., Hewitt, L., Tenenbaum, J. & Solar-Lezama, A.. (2019). Learning to Infer Program Sketches. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:4861-4870 Available from https://proceedings.mlr.press/v97/nye19a.html.

Related Material