College of IST Distinguished Lecture Series - Raymond J. Mooney

4 p.m., to 5 p.m. Thursday, February 25
Join Raymond J. Mooney for a talk in the College of IST Distinguished Lecture Series, "Deep Learning for Automating Software Documentation Maintenance".
Applying deep learning to large open-source software repositories offers the potential to develop many useful tools for aiding software development, including automated program synthesis and documentation generation. Specifically, we have developed methods that learn to automatically update existing natural language comments based on changes to the body of code they accompany. Developers frequently forget to update comments when they change code, which is detrimental to the software development cycle, causing confusion and bugs. First, we use methods for "just in time" comment/code inconsistency
detection which learn to recognize when changes to code render it incompatible with its existing documentation. We then learn a model that appropriately updates a comment when it is judged to be inconsistent. Our approach learns to correlate changes across two distinct language representations,  generating a sequence of edits that are applied to an existing comment to reflect source code modifications. We train and evaluate our model using a large dataset collected from commit
histories of open-source Java software projects, with each example consisting of an update to a method and any concurrent edit to its corresponding comment. We compare our approach against multiple baselines using both automatic metrics and human evaluation. Results reflect the challenge of this task and that our model outperforms many baselines with respect to detecting inconsistent comments and appropriately updating them.
Dr. Mooney is a professor in the Department of Computer Science at the University of Texas at Austin. He is an author of over 180 published research papers, primarily in the areas of machine learning to natural language processing. He was the president of the International Machine Learning Society from 2008-2011, program co-chair for AAAI 2006, general chair for HLT-EMNLP 2005, and co-chair for ICML 1990. He is a Fellow of AAAI, ACM, and ACL and the recipient of the Classic Paper award from AAAI-19 and best paper awards from AAAI-96, KDD-04, ICML-05, and ACL-07.