BAPP #7: Fixing Healthcare with Bayesian Inference and Stan

• What we'll do
Time: Starts 6:00 with drinks, food, and conversation. Presentation starts at 6:30pm, ends at 8PM. Style: Talk

Presenter: Eric Novik, CEO and cofounder. Daniel Lee, CTO and cofounder.

Links: www.generable.com, https://www.generable.com/2018/01/04/san-francisco.html

Stan is an open source probabilistic programming language that was released in 2011 by a small team of developers at Columbia University. Today, Stan has thousands of users across many industries and is growing rapidly. Stan includes a flexible modeling language for expressing statistical models and a set of inference algorithms, most notably HMC with NUTS which is revolutionizing high dimensional statistical computing. Stan also includes a set of ODE (Ordinary Differential Equation) solvers developed in collaboration with Novartis that are particularly useful in Pharma and other industries that are interested in mechanistic and semi-mechanistic models. At Generable we are building an analytics platform targeted at the healthcare market where data alone are not enough to inform statistical inferences. This is particularly relevant to Rare Diseases, Pediatrics, and Oncology where patient populations are small and the mechanism of action may not be understood but could be reasonably approximated. In these cases, most traditional methods of inference simply do not work.

In this talk will focus on a clinical meta-analysis model and highlight all the steps of Bayesian analysis, emphasizing which parts of the workflow we are automating.

• What to bring

• Important to know