DrThin Nguyen

My research advances causal AI: machine learning that goes beyond correlation to reason about causes, mechanisms, interventions and explanations. The central vision is to build systems that ask not only what is likely to happen, but why it happens, what would change under intervention, and when a learned model can be trusted under distribution shift. To this end I develop methods for causal discovery, causal inference, trustworthy and explainable AI, and causal reasoning in foundation models, with contributions including acyclicity-free and Bayesian causal discovery, conditional-independence testing with neural density models, robust causal direction and noise-model estimation, root-cause explanation, causal domain generalisation, and causal alignment and steering of large language models.