Beschrijving
In this episode we sit down with PhD researcher Dennis Thumm to explore his work in causal and probabilistic machine learning. Taking safeguarding to a new scale, we discuss how AI could protect whole populations at once, from predicting natural disasters to testing drugs without risking human or animal lives.
Beyond the research, we consider synthetic data, unobserved confounders, spurious correlations and the human biases that find their way into the models we build. If AI is only as flawed as its creators, how do we make it safer, fairer and more transparent, and how far should we trust…

