Probabilistic Soft Logic

Abstract
There is a growing interest in methods for exploiting causal or correlational dependencies in structured domains. Exploiting such dependencies often results in improved predictive performance on complex inference tasks in diverse domains such as information integration, natural language processing, and computer vision. In this presentation, we introduce probabilistic soft logic (PSL), a general-purpose framework for expressing, reasoning about and learning structural dependencies.
PSL provides a declarative language tailored to relational domains that require reasoning about similarity and/or probability. Some of the novel aspects of PSL include a representation based on continuous valued random variables, efficient polynomial-time inference algorithms, native support for reasoning about sets, and the ability to estimate confidences values for predictions. This presentation provides a detailed account of PSL covering its mathematical foundation, logic programming semantics, inference and learning algorithms, scalability through parallelization, and different applications. We close with a demonstration of the PSL system implementation.

Presentation