My talk at University College London this Monday

It’s the Gatsby seminar in the Computational Neuroscience Unit at University College London, Mon 18 Jan at 4pm:

Creating structured and flexible models: some open problems

A challenge in statistics is to construct models that are structured enough to be able to learn from data but not be so strong as to overwhelm the data. We introduce the concept of “weakly informative priors” which contain important information but less than may be available for the given problem at hand. We also discuss some related problems in developing general models for taxonomies and deep interactions. We consider how these ideas apply to problems in social science and public health. If you don’t walk out of this talk a Bayesian, I’ll eat my hat.

P.S. Link updated.

2 thoughts on “My talk at University College London this Monday

  1. Link seems to be broken.

    As it is being fixed, I was not able to check my old course notes from Ian Hacking class, but his take on Peirce's Faith, Hope and Charity justification of statistical methods may nicely fit with "weakly informative priors"

    If I am recalling roughly right

    Faith – was the enquiring communities would persist in their work

    Hope – that reality was fairly smooth – not a lot of interactions especially not with time (Hume's concern)

    Charity – that individual investigators should forgoe their prior information and be charitable to other investigators and not use it.

    Perhaps using just a bit of the available prior information would be charitable enough

    (think I got at least two mixed up though)

    Keith

Comments are closed.