"The Blessings of Multiple Causes" (Next Week at the Statistics Seminar)
conservation notice: Announcement of a highly technical talk, in a
subject you're not interested in, at a university you're not near.
I am very much looking forward to our first talk of the semester, which I am sure will evoke lively discussion:
- Yixing Wang, "The Blessings of Multiple Causes"
- Abstract: Causal inference from observational data is a vital
problem, but it comes with strong assumptions. Most methods assume that we
observe all confounders, variables that affect both the causal variables and
the outcome variables. But whether we have observed all confounders is a
famously untestable assumption. We describe the deconfounder, a way to do
causal inference from observational data allowing for unobserved
- How does the deconfounder work? The deconfounder is designed for problems of
multiple causal inferences: scientific studies that involve many causes whose
effects are simultaneously of interest. The deconfounder uses the correlation
among causes as evidence for unobserved confounders, combining unsupervised
machine learning and predictive model checking to perform causal inference. We
study the theoretical requirements for the deconfounder to provide unbiased
causal estimates, along with its limitations and tradeoffs. We demonstrate the
deconfounder on real-world data and simulation studies.
- Time and place: 4--5 pm on Monday, 3 February 2020, in Hamerschlag Hall B103
As always, talks are free and open to the public.
Enigmas of Chance;
Constant Conjunction Necessary Connexion
Posted at January 28, 2020 12:06 | permanent link