Visualizing the Uncertainties of Probabilistic Soft Logic With a Graphical Model

Jason Ting

University of California, Santa Cruz

CMPS 261, Winter 2019

Probabilistic Soft Logic is a machine learning framework used in the Statistical Relational Learn- ing, or SRL, community that generates a probabilistic model to find soft truth. Unlike traditional machine learn- ing with IID methods, Probabilistic Soft Logic breaks the IID assumption by utilizes collective classification to make an inference based on the model. Furthermore, the difference between this probabilistic model and other probabilistic models like Markov Logic Network, Probabilistic Soft Logic uses soft truth value to make it’s inference. The objective of this paper is to construct a graphical model to visualize the uncertainties and how it impacts the model as a whole. This will be done through an implementation of a multimodal graph with many different nodes and edges that each have different characteristics.

Graphical Representation clicking on node


Source Code

Technical Paper