What can we learn from data and how much learning can we automate? AI for causal inference in health research
Abstract:
The tools referred to as AI may assist, or eventually replace, health researchers who learn from data. To discuss the potential role of AI, this talk describes a taxonomy of learning tasks in science and explores the relationship between two of them: prediction (pattern recognition) and counterfactual prediction (causal inference). AI tools developed for prediction can be helpful for counterfactual prediction when data are adequately combined with causal models developed from expert knowledge. This raises questions about the origin of causal models and about whether AI tools can partly automate causal inference research in the health sciences.
Miguel Hernán uses health data and causal inference methods to learn what works. As Director of the CAUSALab at Harvard, he and his collaborators repurpose real world data into scientific evidence for the prevention and treatment of infectious diseases, cancer, cardiovascular disease, and mental illness. Miguel is co-director of the Laboratory for Early Psychosis (LEAP) Center, principal investigator of the HIV-CAUSAL Collaboration, and co-director of the VA-CAUSAL Methods Core. As the Kolokotrones Professor of Biostatistics and Epidemiology, he teaches at the Harvard T.H. Chan School of Public Health and at the Harvard-MIT Division of Health Sciences and Technology. His free online course “Causal Diagrams” and book “Causal Inference: What If”, co-authored with James Robins, are widely used for the training of researchers.
This seminar will be held in person and streamed online via Zoom. It will not be recorded.
No registration is required.
Thursday 30 November
9:30-10:30am AEDT
Ella Latham Auditorium
Ground floor
Royal Children's Hospital, Melbourne
Or join via Zoom
Go to https://monash.zoom.us/join and enter
Meeting ID: 829 1898 3473
Passcode: 422205