Designing a rational process for hybrid
probabilistic decision making
This project is defined within the scope of the Hybrid Intelligence: Augmenting Human Intellect project. It is a collaboration between partners from Utrecht University and the University of Groningen.
What is our aim?Rational reasoning with evidence is foundational to many fields. One way of reasoning with evidence is through the use of Bayesian Networks. Due to this method's grounding in probability theory and hence, clear rules for representing evidence strength and updating on new information, these Bayesian Networks are a promising tool for rational decision making.
However, Bayesian Networks are very specific in what they require in order to work correctly. Bringing them from an idealised world of pure statistics and probability theory, to the real world of missing data, counterfactual reasoning and human biases, brings about significant problems. The aim of this project is to design simulation-based processes for creating and evaluating Bayesian Networks that can be applied to real-world problems, such as the problem of the reference class, problem of the priors, causality, and others - or to show limitations of such an approach. The process should be hybrid, in that it plays to the strengths of both the BN and human modellers and users (arguers) by combining probabilistic updating rules with rational deliberation to decide which events, causalities and probabilities to apply.
Why is this important?The research is relevant as a contribution to rationally reasoning with evidence, to understanding how these processes might actually work, and to supporting such reasoning by human-machine collaboration. We investigate and develop methods for the collaboration between humans and machines combining knowledge-based and probabilistic reasoning, thereby augmenting what either can do by themselves.
How will we approach this?We build agent-based simulations as ground truth for Bayesian network models, develop hybrid collaboration mechanisms, and investigate the results in case studies, in particular crime investigation.