Description
Date depot: 18 octobre 2021
Titre: Security Analysis of Source Code Using Machine Learning
Directeur de thèse:
Pietro MICHIARDI (Eurecom)
Encadrant :
Antonino SABETTA (SAP)
Domaine scientifique: Sciences et technologies de l'information et de la communication
Thématique CNRS : Intelligence artificielle
Resumé: This document describes the challenges and a research plan for to the investigation of novel approaches to the analysis of software security atthe level of source code using machine learning techniques. This research plan will be executed by Mr Niccolo Togni for the doctorate he will pursue under the supervision of Eurecom and SAP.
The scope of this research includes a study of existing methods for software security analysis at the source-code level, and an investigation of novel ways to represent source code in a way that is amenable to being used in combination with cutting-edge machine-learning methods. The thesis will study different scenarios, starting from the detection of source code changes that introduce xes to security defects, and proceeding to extend the scope to detect source code changes that introduce security-relevant defects.
The evaluation of the methods proposed will be carried out using publicly available datasets, and, if appropriate, also using SAP-owned data. These datasets will be extended or enriched as needed for the purposes of thesis.
Eurecom will ensure the academic supervision and the scienti c rigour of the methods employed during this research, as well as the soundness of its results. SAP will contribute by framing the research with real-world requirements, stemming from industrial needs and hands-on practice; SAP will be an early adopter of the research results generated during this thesis.
Doctorant.e: Togni Niccolò