Defence of doctoral thesis in the field of computer science, M.Sc. Pedram Daee

Title of the doctoral thesis is "Probabilistic user modelling methods for improving human-in-the-loop machine learning for prediction"
CS_defence_3 photo by Matti Ahlgren

Whether it is a user searching for information in a search system or a doctor working with a cancer diagnostic system, humans and intelligent systems are increasingly interacting with each other. In many of these applications, human involvement in the form of data provider or expert of the task is crucial. However, human communication with intelligent systems is usually constrained by (i) the interaction channels, i.e., how human knowledge can be applied in the system, and (ii) the interaction budget, i.e., how much the user is willing to interact with the system. This dissertation presents new methods, based on probabilistic machine learning, to improve these constraints.

The dissertation investigates topics in human–computer interaction and interactive machine learning, with more emphasize on the latter. In particular, the work has contributed and borrowed ideas from cognitive science and computational rationality, user interface design, probabilistic inference, and active learning. The core aim of the thesis is to enable the intelligent system to maintain a model of the user to better explain the user behavior and to optimize the interaction. This user model can help a recommender system, such as streaming services like Netflix, to detect the interest of a user after few recommendation rounds or to help a doctor assistant system to ask the most informative question about a patient.

Opponent: Professor Roderick Murray-Smith, University of Glasgow, Scotland

Custos: Professor Samuel Kaski, Aalto University School of Science, Department of Computer Science

Contact information of the doctoral student: Pedram Daee, [email protected], +358 504303208

The public defence will be organised via Zoom. Link to the event

Zoom Quick Guide

The doctoral thesis is publicly displayed 10 days before the defence as online display in the Aaltodoc publication archive.

  • Published:
  • Updated: