Probabilistic soft logic (PSL) is a machine learning framework for developing probabilistic models. PSL models are easy and fast: you can define them using a straightforward logical syntax and solve them with fast convex optimization. PSL has produced state-of-the-art results in many areas spanning natural language processing, social-network analysis, and computer vision. The PSL framework is available as an Apache-licensed, open source project on GitHub with an active user group for support. Ready to get started with PSL? Jump to tutorials, key papers, modeling guides, support forum, or source code.

Key Papers

OverviewHinge-loss Markov Random Fields and Probabilistic Soft Logic
Weight LearningHinge-loss Markov Random Fields: Convex Inference for Structured Prediction
Online InferenceBudgeted Online Collective Inference
Latent VariablesPaired-Dual Learning for Fast Training of Latent Variable Hinge-Loss MRFs
Topic ModelingLatent Topic Networks: A Versatile Probabilistic Programming Framework for Topic Models

See the publications page for more technical information and applications.


We recorded a series of introductory classes on PSL:

The following slides provides more information on PSL:

PSL Overview Slides

Getting Started with PSL models

Our Getting Started guide can help you start writing PSL models no matter what your background. We have a simple command-line interface for those without a programming background, a Groovy interface for more powerful models, and a Java library for experienced hackers.

Getting Help with PSL

Stuck? Have a question? We have an active mailing list for discussing questions and problems with PSL.