Talk Title
Generative Modeling
Talk Summary
Over the last few years the term “generative modeling” has become increasingly popular. The precise definition of the term, however, tends to be vague and often inconsistent across the various fields where it is used. In this talk, I will review some explicit definitions of generative modeling, with an emphasis on _narratively_ generative modeling as a tool for using implicit domain expertise to motivate sophisticated probabilistic models and its connections to data fusion and causal inference.
Speaker’s Bio
Michael Betancourt is the principal research scientist at Symplectomorphic, LLC where he develops theoretical and methodological tools to support practical Bayesian inference. In addition to hosting courses on Bayesian modeling and inference he also collaborates and consults on analyses in epidemiology, pharmacology, engineering, ecology, marketing, and physics, amongst others. Before moving into statistics Michael earned a B.S. from the California Institute of Technology and a Ph.D. from the Massachusetts Institute of Technology, both in physics.
This event is co-organised by I-X, the Department of Mathematics, and Machine Learning & Global Health Network.
Please note that the event is open to Imperial staff and students. External participants are welcome to join online via Teams.