[FOM] Postdoc at U of Toronto

[Kenny Easwaran]

https://academicjobsonline.org/ajo/jobs/14556

Area of Research: Mathematical Logic and Mathematical Statistics

Description of duties: Conduct collaborative research outlined in an
awarded NFRFE-2018 New Frontier in Research proposal, titled
“Rebuilding the Foundations of Bayesian Statistics on Nonstandard
Analysis”.

Our proposal aims to rebuild the mathematical foundations of Bayesian
statistics using tools from mathematical logic and nonstandard
analysis that were unavailable at the time Bayesian analysis was
formalized. Our goal is to tackle one of the grand challenges in
mathematical statistics: fully characterize the relationship between
Bayesian and frequentist inference, the two dominant approaches to
statistical inference. In recent work, replacing the standard model of
countably additive probability theory with a nonstandard model allowed
us to settle a longstanding open problem of deriving a Bayesian
characterization of frequentist extended admissibility. We believe
that rapid progress on the grand challenge can be achieved by
committing additional resources to this unconventional approach.

The candidate may also be asked to teach a small course related to the
research, although the primary duty is research.

The Principal Investigator is Daniel M. Roy (Statistics/CS; Toronto).
The following researchers are co-applicants on the funded project and
will be involved in research and informal advising: William Weiss
(Math; Toronto), Robert Anderson (Math, Econ; Berkeley), Nancy Reid
(Statistics; Toronto), and Kenny Easwaran (Philosophy: Texas A&M).
Collaborators include Aaron Smith (Probability; U. Ottawa) and Haosui
Duanmu (Econ; Berkeley).

Salary: $45,000 to $65,000 CAD per year, based on experience; negotiable.

Required qualifications:

Candidates should have obtained or expect to obtain a Ph.D. (or
equivalent) in statistics, mathematics, or a related field by
September 2019. Candidates should have a strong background in at least
one of mathematical logic, probability theory, or
mathematical/theoretical statistics and should have demonstrated
promise in research as evidenced by peer-reviewed research.