Jonas De Maeyer

Investigated path planning in dynamic environments using multi-agent reinforcement learning (MARL). Developed a hierarchical environment-splitting framework to enable efficient regional retraining.

Investigated several explanation methods for NLP models (LIME, SHAP, DeepLIFT, Integrated Gradients). Developed an evaluation framework with six criteria assessing properties such as faithfulness, consistency, and compactness.

Master of Science (M.S.) in Computer Science: Data Science and AI graduated with distinction (75%).

Bachelor of Science (B.S.) in Computer Science graduated with highest distinction (85%).