Agents for Software Development
Graham Neubig (Carnegie Mellon University)
Software is one of the most powerful tools that we humans have at our disposal; it allows a skilled programmer to interact with the world in complex and profound ways. However, at the same time software systems are complex, fragile, and even dangerous. Can we develop AI agents that help us develop real-world software, particularly in the context of real-world software development tasks in large software repositories, in all their complexity? In this tutorial I will discuss the state-of-the-art in software development agents, including challenges with respect to identifying which files to edit, how to edit them, how to test edits and recover, and how to train and evaluate models. In addition, I will address some challenges beyond simple writing code, such as how to process multimodal data, how to combine web browsing with coding, and how to perform data science tasks with software development models. I will provide examples from OpenHands, an open-source toolkit that implements many of the methods that I discuss: https://github.com/All-Hands-AI/OpenHands
Harmonized Coding with AI: LLMs for Qualitative Analysis in Software Engineering Research
Christoph Treude (Singapore Management University), Youmei Fan (Nara Institute of Science and Technology), Tao Xiao (Nara Institute of Science and Technology), and Hideaki Hata (Shinshu University)
Qualitative bottom-up coding is essential for identifying themes and patterns in complex data. This tutorial demonstrates how LLMs such as ChatGPT can support the qualitative coding process for software engineering research. Participants will walk through an entire coding exercise, learning to identify themes through open coding, consolidate themes by refining and merging codes, and conduct inter-rater agreement by standardizing codebooks and testing agreement between human and AI coders. Using hands-on exercises and real-world examples, this session highlights effective human-AI collaboration and strategies to ensure transparency, trustworthiness, and methodological rigor. Designed for researchers at all levels of experience, this tutorial equips participants with practical techniques to analyze software engineering data.
Accepted Papers
| Title | |
|---|---|
| Agents for Software Development Tutorials | |
| Harmonized Coding with AI: LLMs for Qualitative Analysis in Software Engineering Research Tutorials |
Call for Tutorials
The MSR Tutorials track aims to invite experienced software repository miners to give informative tutorials to our broad community, whether newcomers or mining experts. These tutorials will cover various topics related to mining software repositories.
We encourage researchers to submit an abstract outlining a talk, with a maximum length of one page (plus up to one additional page of references). We are soliciting abstracts in two categories:
- Research Problem: the talk outlines a single problem in an academic/industrial context that could be addressed using data from software repositories.
- Lessons Learnt & Opportunities: the talk outlines a series of actionable implications and provides promising directions for future researchers.
If you have any questions, please contact the co-chairs: Dong Wang (d.wang@ait.kyushu-u.ac.jp) and Ayushi Rastogi (a.rastogi@rug.nl)
Submission
A one-page PDF file is required for submission, including the title, speaker names, affiliations, and the outline of the tutorial talk. Additionally, up to one page of references can be included. We do not require any specific template for the submission stage.
If your submission is accepted, you can have your abstract published in the conference proceedings. In this case, your abstract must conform to the official “ACM Primary Article Template”, which can be obtained from the ACM Proceedings Template page. If you are using LaTeX, please use the “sigconf” option, as well as the “review” option, to produce line numbers for easy reference by the reviewers. To do so, please include the following LaTeX code at the start of your document:
\documentclass\[sigconf,review,anonymous]{acmart}
\acmConference\[ICSE 2024]{46th International Conference on Software Engineering}{April 2024}{Lisbon, Portugal}
If you would like to submit your tutorial for consideration for the 2024 MSR Tutorials, please submit it at https://msr24tutorial.hotcrp.com/
Evaluation Criteria
The submissions will be evaluated by the two tutorial track co-chairs. We will assess the relevance and clarity of the problem description, the lessons learned, and the research opportunities, along with the context, in submissions from researchers.
Accepted Papers
- The official publication date is when the proceedings become available in the ACM or IEEE Digital Libraries. Note that this date may be up to two weeks before the first day of ICSE 2024. It’s important to remember that the official publication date may affect the deadline for any patent filings related to published work.
- Purchases of additional pages in the proceedings are not allowed.
Important Dates (all dates are in AoE)
- Tue 19th of December 2023 (Proposal submission deadline),
- Thu 21st of December 2023 (Author Notification),
- Sun 28th of January 2024 (Camera-ready version)