Current PhD students

In parentheses, the percentage of my implication in each project.

(50%) Andrés LOPEZ who works in a thesis entitled “A tooled method to design and create AI-intensive software product lines” co-supervised with Assoc. Prof. Paola VALLEJO (Univ. EAFIT, Colombia) and Assoc. Prof. Viviana COBALEDA (Universidad de Antioquia, Colombia) .

Estimated period: January 2024 - December 2027.

In this project, we want to study the possibility of software product lines that integrate ML-based components. The project will seek answers to the following questions: do software systems with AI/ML components have particular needs with respect to variability modeling? Which of these needs are not covered by current languages for modeling variability and how to cover them? How could the reusability and configurability needs of AI/ML components be covered in order to derive diverse and varied products from them? How to evaluate the proposed method and what are the results and significance of the evaluation results?

(50%) Hiba HNAINI who works in a thesis entitled “Towards a unifying framework for the specification, formalization, and analysis of secure hardware and software architectures” co-supervised with Assoc. Prof. Joel CHAMPEAU (ENSTA Bretagne, France) and Assoc. Prof. Paola VALLEJO (Univ. EAFIT, Colombia).

Estimated period: March 2021 - March 2024.

The main goal of this project is to design and evaluate a new multi-paradigm security modeling approach and its engineering framework (engineering process and tooling) allowing the specification, formalization and analysis of secure hardware and software architectures.

(80%) Anne-Laure WOZNIAK who works in a CIFRE thesis entitled “Innovative framework for testing neural network-based software components for critical systems” co-supervised with Assoc. Prof. Dorian CAZAU (ENSTA Bretagne, France).

Estimated period: February 2021 - February 2024.

This project seeks to address the problems related to behavioural anomalies in neural network-based components used in safety-critical systems. In particular we intend to propos a generic method to guide the robustness testing of ML-based systems. The main research question that this work addresses is thus: How can the robustness of ML-based systems be assessed in practice?

(50%) Camilo CORREA who works in a thesis entitled “CLEVER, a Compositional, Learnable, Explainable, Verifiable, Executable, Relational, Recursive and Reifying formalism to represent and simulate Intelligent Self-Adaptive Systems” co-supervised with Jacques ROBIN (Univ. Paris 1, France).

Estimated period: October 2020 - October 2023.

This project aims to develop a framework for the design and implementation of intelligent self-adaptive systems based on the convergence of technological developments in the fields of symbolic and statistical AI.