Learning Approximately Objective Priors
نویسندگان
چکیده
Informative Bayesian priors are often difficult to elicit, and when this is the case, modelers usually turn to noninformative or objective priors. However, objective priors such as the Jeffreys and reference priors are not tractable to derive for many models of interest. We address this issue by proposing techniques for learning reference prior approximations: we select a parametric family and optimize a black-box lower bound on the reference prior objective to find the member of the family that serves as a good approximation. We experimentally demonstrate the method’s effectiveness by recovering Jeffreys priors and learning the Variational Autoencoder’s reference prior.
منابع مشابه
Variational Reference Priors
Posterior distributions are useful for a broad range of tasks in machine learning ranging from model selection to reinforcement learning. Given that modern machine learning models can have millions of parameters, selecting an informative prior is typically infeasible, resulting in widespread use of priors that avoid strong assumptions. For example, recent work on deep generative models (Kingma ...
متن کاملData-dependent PAC-Bayes priors via differential privacy
The Probably Approximately Correct (PAC) Bayes framework (McAllester, 1999) can incorporate knowledge about the learning algorithm and data distribution through the use of distributiondependent priors, yielding tighter generalization bounds on data-dependent posteriors. Using this flexibility, however, is difficult, especially when the data distribution is presumed to be unknown. We show how an...
متن کاملSample-Efficient Reinforcement Learning through Transfer and Architectural Priors
Recent work in deep reinforcement learning has allowed algorithms to learn complex tasks such as Atari 2600 games just from the reward provided by the game, but these algorithms presently require millions of training steps in order to learn, making them approximately five orders of magnitude slower than humans. One reason for this is that humans build robust shared representations that are appl...
متن کاملLocation Reparameterization and Default Priors for Statistical Analysis
This paper develops default priors for Bayesian analysis that reproduce familiar frequentist and Bayesian analyses for models that are exponential or location. For the vector parameter case there is an information adjustment that avoids the Bayesian marginalization paradoxes and properly targets the prior on the parameter of interest thus adjusting for any complicating nonlinearity the details ...
متن کاملExploiting Informative Priors for Bayesian Classification and Regression Trees
A general method for defining informative priors on statistical models is presented and applied specifically to the space of classification and regression trees. A Bayesian approach to learning such models from data is taken, with the MetropolisHastings algorithm being used to approximately sample from the posterior. By only using proposal distributions closely tied to the prior, acceptance pro...
متن کامل