I seek to discover relationships within data via synthesizing BLOG models to describe them. BLOG (Bayesian LOGic) is a first-order probabilistic programming language that details probability distributions over worlds containing sets of objects. This sort of learning has previously been done with Probabilistic Relational Models, which are a highly restricted special case of BLOG models. I synthesize programs using a local search algorithm that maximizes over the likelihood of a model with the given data while penalizing complexity. I apply the algorithm to learning the generative model describing how parts of citations are written and show that it is capable of learning accurate and useful relationships.
Title
Inductive Program Synthesis in BLOG
Published
2018-08-09
Full Collection Name
Electrical Engineering & Computer Sciences Technical Reports
Other Identifiers
EECS-2018-107
Type
Text
Extent
36 p
Archive
The Engineering Library
Usage Statement
Researchers may make free and open use of the UC Berkeley Library’s digitized public domain materials. However, some materials in our online collections may be protected by U.S. copyright law (Title 17, U.S.C.). Use or reproduction of materials protected by copyright beyond that allowed by fair use (Title 17, U.S.C. § 107) requires permission from the copyright owners. The use or reproduction of some materials may also be restricted by terms of University of California gift or purchase agreements, privacy and publicity rights, or trademark law. Responsibility for determining rights status and permissibility of any use or reproduction rests exclusively with the researcher. To learn more or make inquiries, please see our permissions policies (https://www.lib.berkeley.edu/about/permissions-policies).