Multi-Entity Bayesian Networks Learning for Predictive Situation Awareness
Date
2017
Authors
Park, Cheol Young
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Over the past few decades, machine learning has led to substantial changes in Data Fusion Systems throughout the world. One of the most important application areas for data fusion is to support situation awareness for command and control. Situation Awareness (SAW) is perception of elements in the environment, comprehension of the current situation, and projection of future status. Predictive Situation Awareness (PSAW) emphasizes the ability to make predictions about aspects of a temporally evolving situation. PSAW requires a semantically rich representation to handle a complex real world situation and ability to reason under uncertainty about the situation. Probabilistic ontologies are able to address the requirements of PSAW, by augmenting standard ontologies with support for uncertainty management. PR-OWL (Probabilistic Web Ontology Language), a representation language for probabilistic ontologies, is founded on Multi-Entity Bayesian Networks (MEBN). MEBN combines First-Order Logic with Bayesian Networks for representing and reasoning about uncertainty in complex, knowledge-rich domains. MEBN goes beyond standard Bayesian networks to enable reasoning about an unknown number of entities interacting with each other in various types of relationships, a key requirement for PSAW. MEBN models have heretofore been constructed manually by a domain expert. However, manual MEBN modeling is labor-intensive and insufficiently agile. To address these problems, an efficient method is needed for MEBN modeling. One of the methods is to use machine learning to learn a probabilistic ontology in whole or in part from data. In the era of Big Data, data-rich environments, characterized by uncertainty and complexity, have become ubiquitous. The larger the data sample is, the more accurate the results of the machine learning approach can be. Therefore, machine learning has potential to improve the quality of MEBN models. In this research, we study a machine learning method from data for MEBN-based probabilistic ontologies for PSAW. Specifically, we introduce a MEBN learning framework to develop a MEBN model from a combination of domain expert's knowledge and data. To support the framework, we present a bridge model between a MEBN model and a relational model, a reference model supporting the design of a MEBN model for PSAW, and a parameter learning algorithm given a MEBN model. The presented methodology is evaluated on three example use cases: (1) a Critical Infrastructure Defense system, (2) a Maritime Domain Awareness system, (3) a Smart Manufacturing system. Finally, we conduct an experiment to compare the MEBN learning framework and the manual MEBN modeling in terms of development efficiency.
Description
Keywords
Information technology, Computer science, Human-added Machine Learning, Machine learning, Multi-Entity Bayesian Networks, Predictive Situation Awareness, Situation Awareness, Smart System