SI Seminar by Matteo Biagiola
Date: 3 November 2022 / 16:30 - 17:30
USI Campus Est, room D0.03, Sector D // online on MS Teams
You can join here
Reinforcement Learning for Software Testing
Speaker: Matteo Biagiola
Reinforcement Learning (RL) is a learning paradigm which is used to address sequential decision-making tasks. The objective of an RL algorithm is to learn how to behave in an environment to solve a particular task. RL has regained popularity in recent years, thanks to advances in deep learning that made it applicable to complex domains. RL is also being applied to software engineering tasks, achieving state-of-the-art performance in most cases. In this talk, Matteo will describe the main steps required to model a problem as an RL problem. Then, he will show an application of RL for a software testing task, i.e., test case prioritization, describing the different ways such problem can be modelled and the impact of the modelling on both prioritization effectiveness and efficiency.
Matteo is a postdoctoral fellow at the Software Institute (USI) in Lugano. He obtained his PhD in 2020 from Università di Genova in a joint program with Fondazione Bruno Kessler (FBK) in Trento, under the supervision of Prof. Paolo Tonella and Prof. Filippo Ricca. He is interested in software testing, with a particular focus on test case generation for web applications and reinforcement learning systems. He serves on the program committees of software engineering conferences such as ICST and SSBSE and review for several software engineering journals, including TOSEM, TSE, EMSE, JSS and IST.
Chair: Luca Chiodini
ℹ️ The seminar will be in presence for everyone in room D0.03. If you are unable to attend in presence, the SI website contains a link to a Teams video call in the “location" field (i.e., click on “D0.03").
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.