Texte à Master 2 Internship Subject -

Title: Neurosymbolic Test Case Generation and Selection Supervisors: Jérémie Guiochet (https://homepages.laas.fr/guiochet), Hélène Waeselynck (https://homepages.laas.fr/waeselyn/)

Profile: Computer Science Level: Master 2 (MSc)

Duration: 5–6 months

Location: LAAS-CNRS, Toulouse, France (https://www.laas.fr)

Salary: ~€600/month

Test case generation and selection are major challenges in software testing, particularly for autonomous systems such as vehicles and robots, which are often evaluated using high-fidelity simulators. TAF (Testing Automation Framework), developed at LAAS-CNRS (https://wp.laas.fr/taf), generates diverse test cases from XML specifications containing parameters and constraints. It first produces logical test descriptions, then converts them into concrete test cases that can be executed by the system under test (SUT). Currently, TAF focuses on generation. However, executing large numbers of tests can be costly particularly when a high-fidelity simulator in used. Machine learning (ML) could help prioritize tests that are more likely to reveal defects or cover testing objectives, reducing execution costs while improving testing efficiency. The objective of this internship is to investigate how ML can guide test case selection and prioritization within a TAF-based pipeline. The approach will combine constraint-based generation, which ensures the validity of test descriptions, with data-driven predictions of their potential to reveal defects or to achieve test objectives. The internship should deliver a prototype integrating ML-based prioritization into TAF, an experimental dataset, and an evaluation of its benefits and limitations. The internship is suited to a Master's student in Computer Science interested in machine learning, software testing, or dependable autonomous systems. Python and basic ML knowledge are expected; familiarity with constraint solving or simulation-based testing is an advantage.

Application : CV + cover letter + transcripts of records for the last two academic years