Freer, CE and Roy, DM (2010) Posterior distributions are computable from predictive distributions. Journal of Machine Learning Research, 9. pp. 233-240. ISSN 1532-4435Full text not available from this repository.
As we devise more complicated prior distributions, will inference algorithms keep up? We highlight a negative result in computable probability theory by Ackerman, Freer, and Roy (2010) that shows that there exist computable priors with noncomputable posteriors. In addition to providing a brief survey of computable probability theory geared towards the A.I. and statistics community, we give a new result characterizing when conditioning is computable in the setting of exchangeable sequences, and provide a computational perspective on work by Orbanz (2010) on conjugate nonparametric models. In particular, using a computable extension of de Finetti's theorem (Freer and Roy 2009), we describe how to transform a posterior predictive rule for generating an exchangeable sequence into an algorithm for computing the posterior distribution of the directing random measure. Copyright 2010 by the authors.
|Divisions:||Div F > Computational and Biological Learning|
|Depositing User:||Cron job|
|Date Deposited:||04 Feb 2015 23:01|
|Last Modified:||05 Feb 2015 06:29|