Code Automation Unleashed: A Journey Into Transformative AI for Software Engineering

Abstract: 

In recent years, the application of Artificial Intelligence (AI) to software engineering tasks has been changing how software is developed and maintained. In this seminar, I will explain how AI, and in particular deep learning (DL) models, can streamline complex code-related tasks like those requiring the manipulation of multimodal data, such as code and technical natural language (e.g., automatically documenting a piece of code in natural language). 

I will review crucial technical aspects, including the importance of the pre-training phase that has led to unprecedented automation for specific code-related tasks. Then, I will present AI-based recommender systems developed in the context of my PhD, by discussing the main challenges faced during their design and empirical evaluation. 

Following this, I will conclude the seminar by highlighting future directions for the field as an integral part of my medium-term research agenda.

Bio:

Antonio Mastropaolo is currently a Ph.D. candidate at the Università della Svizzera italiana (USI) in Switzerland, where he contributes to the Software Institute within the Faculty of Informatics. He started his Ph.D. in October 2020, after earning a Master of Science in Software System Security from Università degli Studi del Molise, Italy, in July 2020. 

Antonio's research focuses on software engineering, with a particular interest in automating code-related tasks such as bug-fixing and code summarization, to enhance the quality of software systems. His research exploits techniques which embrace natural language processing (NLP) and artificial intelligence (AI), with a strong focus on deep learning (DL) methods, to automate code-related tasks. Antonio's research has been featured in prominent software engineering forums, including the International Conference on Software Engineering (ICSE), the International Conference on Automated Software Engineering (ASE), and top-tier journals in the field, such as Transaction on Software Engineering (TSE), and Empirical Software Engineering (EMSE).