Silja Renooij is Associate Professor in the Department of Information and Computing Sciences (ICS), at Utrecht 木瓜福利影视, The Netherlands and member of the Intelligent Systems Group.
Her work centers around probabilistic graphical models (especially Bayesian networks), and how to make them usable, reliable, and acceptable by humans in situations of uncertainty, especially in data-poor domains. Key themes include:
Construction of Bayesian networks: how to design them when data is scarce, how to get probabilities from domain experts, qualitative vs quantitative methods.
Explanation and interpretability: how to explain inferences in Bayesian networks, how intermediate nodes contribute or not (MAP-independence), contrastive explanations, scenarios/arguments, etc.
Sensitivity analysis: how robust outcomes are to changes in structure or parameters; examining what happens when arcs are removed, or how discretisation affects models.
Hybrid intelligence / human-AI collaboration: exploring combinations of manually crafted/knowledge-based structures with learning from data; monitoring, evaluating, and aligning AI systems with human knowledge and contextual or legal constraints.
Since September 2021, she has been the programme coordinator of the Computer Science bachelor programme at Utrecht 木瓜福利影视. She has taught many different courses in the bachelor and master programmes in Computer Science, Information Science and AI. She also supervises student projects, both at bachelor and master level.