Probabilistic models for human judgments about uncertainty in intuitive inference tasks

Dr. Philipp Schustek

Thesis supervisor: Dr. Rubén Moreno-Bote

15/06/2018

 

Brief description:

How can we be sure of our knowledge? This is of tremendous importance, for instance, to determine if we can trust a medical diagnosis before undergoing risky surgery.

In my PhD project, I explored key abilities for this kind of questions. I designed psychological experiments to test and computationally model how we arrive at confidence statements. I investigated whether our brains can draw on probabilistic computations for judgments about uncertainty thus contributing to a young research field at the intersection of machine learning and cognitive science. At least in certain situations, and somewhat counter to commonly held beliefs, we appear to be better ‘intuitive statisticians’ than previously thought.

 

Experience as PhD student:

Right from the beginning, the UPF and the Center for Brain and Cognition were highly supportive and welcoming. With the backing of my supervisor, I could take on the challenge of a new approach to testing computational hypotheses in my field. At the same time, the UPF promoted the exchange with other, often international, researchers and disciplines creating a stimulating scientific environment. While substantial challenges had to be overcome along the way, I will look back with satisfaction and gratitude at both the personal and professional experiences I have made here in Barcelona