Exploring Software Naturalness throughNeural Language Models

06/22/2020

∙

by   Luca Buratti, et al.

∙

0

∙

share

The Software Naturalness hypothesis argues that programming languages can be understood through the same techniques used in natural language processing. We explore this hypothesis through the use of a pre-trained transformer-based language model to perform code analysis tasks.