Over the past few decades, researchers have made significant strides in automating software development processes. This evolution has transformed the way software is created, maintained, and enhanced. Recently, the integration of Large Language Models (LLMs) into software development has opened new horizons. Researchers have investigated the potential of LLMs and demonstrated that they provide strong performance gains. These models can understand natural language instructions, generate code snippets, and even identify and fix bugs, thereby streamlining the development process. However, software engineering encompasses more than just coding; it involves the continuous improvement of programs to facilitate software maintenance and evolution. This includes tasks like program repair to fix bugs and feature additions to enhance functionality. Traditional automation tools often fall short in these areas, highlighting the need for more advanced solutions. Inspired by these insights, we have developed a novel automated program repair method called \textit{AutoPR}. AutoPR represents a new generation of AI software engineers, leveraging routing algorithms, in-memory caching, and collaborative agent technologies. Its design addresses the current efficiency bottlenecks and quality issues faced in software development.
Keywords: Automated Software Development;Program Repair;AI Software Engineers;Collaborative Agent Technologies;In-Memory Caching;Routing Algorithms
Abstract:
Primary Area: foundation or frontier models, including LLMs
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 6779
Loading