Our course with Per Ola Kristensson (Cambridge), Nikola Banovic (Michigan), and John Williamson (Glasgow) was accepted to CHI 2019 in Glasgow. It’s a 4 unit course. Abstract below:
This course introduces computational methods in human– computer interaction. Computational interaction methods use computational thinking — abstraction, automation, and analysis — to explain and enhance interaction. This course introduces the theory of practice of computational interaction by teaching Bayesian methods for interaction across four wide areas of interest when designing computationally-driven user interfaces: decoding, adaptation, learning and optimization. The lectures center on hands-on Python programming interleaved with theory and practical examples grounded in problems of wide interest in human-computer interaction.
The course will cover:
Decoding: Principled and robust formulations for proba- bilistically decoding noisy sensor observations into user’s intended actions.
Adaptation: Probabilistic adaptation demonstrated by se- quential Monto Carlo approaches for modeling HCI tasks.
Learning: Using inverse reinforcement learning for learning computational models of human behavior.
Optimization: Using Bayesian optimization to perform human-in-the-loop optimization of a user interface problem.