- Review of the 2014 BayesiaLab User Conference
- Judea Pearl
From Bayesian Networks to Causal and Counterfactual Reasoning - Lionel Jouffe
Causal Analysis with Structural Equation Models and Bayesian Networks - Alex Cosmas
So you can predict the future... But can you change it? - Mick McWilliams
Using Bayesian Networks to Model Key Drivers - David Aebischer
Michael Grimes
Bayesian Belief Network Applications for Supporting Warfighters - Ferit Akova
ProvenCare Lumbar Spine Prediction Model - Jack Y. Chen
Revolutionizing Decision Making: How Analytics Will Take Over the Business - Roman Fomin
Prediction of Overall Team Peformance and Injury in Team Sports - Christina Ray
Intelligence Tradecraft and Bayesian Models - Michael Ryall
When nodes think: using BayesiaLab to analyze decisions in game theoretic settings
2014 BayesiaLab User Conference Presentations
Michael Ryall, Ph.D.
Associate Professor of Strategy and Economics, Rotman Business School, University of Toronto, Canada
When nodes think: using BayesiaLab to analyze decisions in game theoretic settings
Comparing the relative difficulty of social science to that of physics, the Nobel Prize-winning physicist Murray Gell-Mann once said, “Imagine how difficult physics would be if electrons could think.” Influence diagrams, the causal modeling analog to decision trees, are composed of decision, random, and payoff nodes. Software tools such as BayesiaLab provide decision support by solving such diagrams for optimal choice policies. Recalling Gell-Mann, however, in most business settings the key “nodes” in an influence diagram represent other self-interested agents, with free wills and independent objectives — that is, nodes that think. Modeling such agents as unthinking, random nodes creates serious analytical blind spots. Instead, situations with strategic agents should be modeled using interactive influence diagrams, a form of game theoretic analysis. I will explain these ideas and illustrate how to implement them using BayesiaLab.