TCSE logo 
 Sigsoft logo
Sustainability badge

This program is tentative and subject to change.

Fri 2 May 2025 16:30 - 16:45 at 212 - AI for Process

When a Commit Message Generation (CMG) system is integrated into the IDEs and other products at JetBrains, we perform online evaluation based on user acceptance of the generated messages. However, performing online experiments with every change to a CMG system is troublesome, as each iteration affects users and requires time to collect enough statistics. On the other hand, offline evaluation, a prevalent approach in the research literature, facilitates fast experiments but employs automatic metrics that are not guaranteed to represent the preferences of real users. In this work, we describe a novel way we employed to deal with this problem at JetBrains, by leveraging an online metric - the number of edits users introduce before committing the generated messages to the VCS - to select metrics for offline experiments. To support this new type of evaluation, we develop a novel markup collection tool mimicking the real workflow with a CMG system, collect a dataset with 57 pairs consisting of commit messages generated by GPT-4 and their counterparts edited by human experts, and design and verify a way to synthetically extend such a dataset. Then, we use the final dataset of 656 pairs to study how the widely used similarity metrics correlate with the online metric reflecting the real users’ experience.

Our results indicate that edit distance exhibits the highest correlation with the online metric, whereas commonly used similarity metrics such as BLEU and METEOR demonstrate low correlation. This contradicts the previous studies on similarity metrics for CMG, suggesting that user interactions with a CMG system in real-world settings differ significantly from the responses by human labelers within controlled environments. We release all the code and the dataset to support future research in the field: https://jb.gg/cmg-evaluation.

This program is tentative and subject to change.

Fri 2 May

Displayed time zone: Eastern Time (US & Canada) change

16:00 - 17:30
16:00
15m
Talk
OptCD: Optimizing Continuous Development
Demonstrations
Talank Baral George Mason University, Emirhan Oğul Middle East Technical University, Shanto Rahman The University of Texas at Austin, August Shi The University of Texas at Austin, Wing Lam George Mason University
16:15
15m
Talk
LLMs as Evaluators: A Novel Approach to Commit Message Quality Assessment
New Ideas and Emerging Results (NIER)
Abhishek Kumar Indian Institute of Technology, Kharagpur, Sandhya Sankar Indian Institute of Technology, Kharagpur, Sonia Haiduc Florida State University, Partha Pratim Das Indian Institute of Technology, Kharagpur, Partha Pratim Chakrabarti Indian Institute of Technology, Kharagpur
16:30
15m
Talk
Towards Realistic Evaluation of Commit Message Generation by Matching Online and Offline Settings
SE In Practice (SEIP)
Petr Tsvetkov JetBrains Research, Aleksandra Eliseeva JetBrains Research, Danny Dig University of Colorado Boulder, JetBrains Research, Alexander Bezzubov JetBrains, Yaroslav Golubev JetBrains Research, Timofey Bryksin JetBrains Research, Yaroslav Zharov JetBrains Research
Pre-print
16:45
15m
Talk
Enhancing Differential Testing: LLM-Powered Automation in Release Engineering
SE In Practice (SEIP)
Ajay Krishna Vajjala George Mason University, Arun Krishna Vajjala George Mason University, Carmen Badea Microsoft Research, Christian Bird Microsoft Research, Robert DeLine Microsoft Research, Jason Entenmann Microsoft Research, Nicole Forsgren Microsoft Research, Aliaksandr Hramadski Microsoft, Sandeepan Sanyal Microsoft, Oleg Surmachev Microsoft, Thomas Zimmermann University of California, Irvine, Haris Mohammad Microsoft, Jade D'Souza Microsoft, Mikhail Demyanyuk Microsoft
17:00
15m
Talk
How much does AI impact development speed? An enterprise-based randomized controlled trial
SE In Practice (SEIP)
Elise Paradis Google, Inc, Kate Grey Google, Quinn Madison Google, Daye Nam Google, Andrew Macvean Google, Inc., Nan Zhang Google, Ben Ferrari-Church Google, Satish Chandra Google, Inc
17:15
15m
Talk
Using Reinforcement Learning to Sustain the Performance of Version Control Repositories
New Ideas and Emerging Results (NIER)
Shane McIntosh University of Waterloo, Luca Milanesio GerritForge Inc., Antonio Barone GerritForge Inc., Jacek Centkowski GerritForge Inc., Marcin Czech GerritForge Inc., Fabio Ponciroli GerritForge Inc.
:
:
:
: