Department of Biostatistics Seminar/Workshop Series

Knowledge Discovery in Unknown Domains with Machine Learning and Bayesian Networks

Stefan Conrady

Managing Partner, Conrady Applied Science, LLC, Franklin, TN

Wednesday, September 7, 1:30-2:30pm, MRBIII Conference Room 1220

We will utilize use unsupervised and supervised learning algorithms to automatically generate Bayesian networks from high-dimensional datasets, such as we might find in financial markets, biological systems or many other domains. Our objective is to gain both a qualitative and quantitative description of the unknown domain by using Bayesian networks. Their computational efficiency and inherently visual structure make Bayesian networks attractive for exploring and explaining complex problems, especially when no a-priori expert knowledge is available.

In addition to generating human-readable and interpretable structures from data, we want to illustrate how we can immediately use machine-learned Bayesian networks as “computable knowledge” for automated inference and formal reasoning. In the quantitative context, we will also show how Bayesian networks can reliably carry out inference with multiple pieces of uncertain and even conflicting evidence. The inherent ability of Bayesian networks to perform computations under uncertainty makes them highly suitable for a wide range of real-world decision making applications.

http://www.conradyscience.com/