Post-Inference Methods for Scalable Probabilistic Modeling and Sequential Decision Making
Abstract
Probabilistic modeling refers to a set of techniques for modeling data that allows one to specify assumptions about the processes that generate data, incorporate prior beliefs about models, and infer properties of these models given observed data. Benefits include uncertainty quantification, multiple plausible solutions, reduction of overfitting, better performance given small data or large models, and explicit incorporation of a priori knowledge and problem structure. In recent decades, an array of inference algorithms have been developed to estimate these models. This thesis focuses on post-inference methods, which are procedures that can be applied after the completion of standard inference algorithms to allow for increased efficiency, accuracy, or parallelism when learning probabilistic models of big datasets. These methods also allow for scalable computation in distributed or online settings, incorporation of complex prior information, and better use of inference results in downstream tasks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2019
- Accession Number
- AD1167998
Entities
People
- Willie Neiswanger
Organizations
- Carnegie Mellon University