When there’s no room for error
Combining data and probability theory to make decisions with precision
Marek Druzdzel, Associate Professor
In a perfect world, decisions would be made based on statistics and facts. Human bias would be removed from the decision-making process and replaced with objectivity and consistency. Even when faced with numerous variables, the correct decision would be made every single time.
In 2015, Marek Druzdzel founded BayesFusion, LLC., a company based on the revolutionary decision-making software he created more than two decades ago.
Druzdzel leads a research group for faculty and students interested in developing techniques and systems that support decision-making under uncertainty. The software he created in 1995, called GeNIe and its Application Programmer’s Interface, SMILE (Structural Modeling, Inference, and Learning Engine) analyzes how the probability that a theory is true is affected by a new piece of evidence. SMILE is a reasoning engine for graphical models, such as Bayesian networks, influence diagrams, and structural equation models. A second generation of this software was released in 2000, and this newest version includes equation-based models and learning.
There are numerous fields that present challenges to the decision-making process. Take healthcare, for example. Healthcare professionals have used GeNIe and SMILE for over 20 years in a variety of situations; from diagnosing liver disorders, to determining the optimal follow-up schedule for cervical cancer patients, and predicting the success in diseases such as inflammatory bowel disease and acute pulmonary hypertension.
Just as humans have symptoms that indicate health disorders, machines share warning signs before they break. By calculating the probabilities of possible equipment failures, SMILE has been used to create optimal maintenance schedules for diesel locomotives, medical devices, database servers and other machines.
From the diagnosis of leaks in water supply networks to the prognosis of the energy output of wind farms, the applications for this software are endless.
Druzdzel started BayesFusion, LLC after obtaining an exclusive license for GeNIe and SMILE from the University of Pittsburgh. By taking the software to the private sector, he can hire programmers to focus on the implementation and documentation. He plans to continue offering the software to the academic community at no charge.
Professors at universities around the world currently use this software in their classrooms. Druzdzel’s students use GeNIe and SMILE to play with their ideas and expand upon the concepts he teaches in class.
Druzdzel believes it’s essential to provide his students with encouragement and personal attention. That teaching philosophy might have been part of the reason he received the Chancellor’s Distinguished Teaching Award in 2007, which is the highest award the university gives for teaching.
Probabilistic and decision-theoretic methods in decision support systems, user interfaces to decision support systems, interactive and mixed-initiative systems for support of strategic planning within organizations
NIH, NSF, DARPA, AFOSR
“When you teach something that is not too far from your research, you are really on the cutting edge. It’s also really cool when you can show some of the things that are happening in the world.”
Marek Druzdzel is an associate professor at the University of Pittsburgh’s School of Computing and Information. To learn more about his research, click here. http://www.pitt.edu/~druzdzel