1 Towards Code Ware Code Models, Prof. Baishakhi Ray

Towards Code-Aware Code Models About the Speaker: Professor Baishakhi Ray is an associate professor at Columbia University in New York. Her research spans the intersection of software engineering and machine learning. She has been honored with prestigious awards such as the IEEE TCSE Rising Star Award and NSF CAREER Award, and has garnered several distinguished paper awards for her research contributions. About the Talk: In her keynote, Professor Ray delves into the challenges and advancements in applying AI to improve software development processes. She explores the concept of "code-aware" machine learning models that better understand the intricacies of code beyond mere text interpretation. Professor Ray addresses the limitations of current large language models (LLMs) that treat code primarily as text, missing the structural and semantic nuances vital for high-quality software development. Through her research, she demonstrates innovative approaches like code-aware prompting and the integration of symbolic logic to enhance unit test generation, significantly improving path coverage and test pass rates compared to baseline models. Professor Ray's work is pivotal in moving towards AI-driven software engineering tools that are not only efficient but also reliable and trustworthy, making a compelling case for the integration of AI to support the entire software development lifecycle. https://fmse.io/