Customizing deep learning models for code completion tasks

Software Institute

Date: 11 April 2024 / 16:30 - 17:30

USI Campus Est, room D1.13, Sector D

Speaker: Alessandro Giagnorio, Ph.D student, Faculty of Informatics at USI, Switzerland, member of the SoftwarE Analytics Research Team (SEART).


Nowadays, many developers rely on tools such as GitHub Copilot for completing the next code tokens or crafting entire functionalities. These tools have proved to enhance developers’ productivity by speeding up code writing and assisting with repetitive tasks. While they often produce astonishing results, like generating entire functions from a simple natural language description, they are generally trained on a large amount of source code. Hence, at inference time, these tools provide the most likely solution generalizing on their extensive acquired knowledge. In this seminar, we propose an empirical study aimed at understanding whether further training these models on specialized corpora, like code written by developers or an organization, can improve the performance of these models on the targeted developers.


Alessandro Giagnorio is a Ph.D. student in the Faculty of Informatics at the Università della Svizzera italiana (USI), Switzerland, where he is part of the SoftwarE Analytics Research Team (SEART). He received his double Master’s degree in Software System Security (Università degli Studi del Molise, Italy), and Software and Data Engineering (Università della Svizzera italiana, Switzerland) in July 2023. He is currently researching new methodologies for personalizing deep-learning models to support developers on code-related tasks.

Chair: Igor Moreno Santos


In February 2019, the Software Institute started its SI Seminar Series. Every Thursday afternoon, a researcher of the Institute will publicly give a short talk on a software engineering argument of their choice. Examples include, but are not limited to novel interesting papers, seminal papers, personal research overview, discussion of preliminary research ideas, tutorials, and small experiments.

On our YouTube playlist you can watch some of the past seminars. On the SI website you can find more details on the next seminar, the upcoming seminars, and an archive of the past speakers.