Psychological findings show that under uncertainty, there are many situations where humans make irrational decisions (Tversky & Kahneman, 1974). By irrational, we mean decisions that violate the laws of classical probability theory and logic. These decisions have a deep impact in current decision support systems, since they cannot be predicted: if decision systems are based on classical probabilistic models and on the notions of set theory, then the system cannot capture human irrationalities that violate these theories.
Under a psychological point of view, these violations are explained by models that primarily assume a preference for fast, heuristic-based processing and strong computational assumptions about the human mind. In order to overcome these assumptions, a new field has emerged that generalizes the notions of classical probability theory by using the mathematical formalisms and concepts of quantum mechanics. This new field is known as quantum cognition and although it is still in its infancy, many models (referred to as quantum-like) have been proposed that are able to predict and accommodate many paradoxical decisions reported in the literature in flexible and elegant decision models (Busemeyer & Bruza, 2012). A more pioneering and recent work has even attempted to apply these quantum-like models (more specifically a quantum-like Bayesian Network) in real world scenarios, such as credit applications (Moreira et. al, 2018).
These networks use the concept of quantum interference to disturb unobserved nodes in a network during probabilistic inferences. Very generally, one can look at interference in the following way. If we model human thoughts as waves, these waves propagate until a person reaches a decision (which causes the wave to collapse into a final decision). However, in the presence of conflicting thoughts or uncertain information, these waves can crash, producing destructive interferences and leading to different decisions that were not taken into consideration through classical models.
Quantum interference is one of the core concepts in quantum cognition and enables the accommodation and prediction of many paradoxical findings in decision-making.However, these quantum-like Bayesian networks are static models. They cannot take into account cycles, which restricts its application to many decision scenarios, and they cannot take into account the evolution of human decisions through time.
The purpose of this project is to explore the principles of evolution in dynamical quantum-like models and how these models can be extended and applied in real-world scenarios, such as medical decision-making, financial models, economics, etc.
Some of the activities involved in this project are the following:
- Development of a quantum-like Markov Model
- Application of quantum-like Markov Models to predict and accommodate paradoxical human decisions
- Application of quantum-like Markov Models in real world problems, such as medical decision-making, economics or finance
We expect the results of this topic to include:
- Gain an understanding of the main principles of evolution quantum-like dynamical model
- An open source framework publicly available for other researchers to use
- Publication of results in high impact factor journals
You may be able to apply for a research scholarship in our annual scholarship round.
- Quantum cognitive models
- Quantum-Like Bayesian Networks
- Quantum-Like Markov Models
- Dynamical models
- Mathematical modelling
- Bayesian statistics
Contact the supervisor for more information.