DIKU Bits: Brain-computer Interfacing for Cognitive Modeling


Tuukka Ruotsalo, ​Associate professor​ in the Machine Learning Section at the Department of Computer Science. 


What if we could just think or naturally react to information that we encounter and computing systems could learn our preferences and provide us information by sensing our thoughts and reactions? What if interacting with computing systems would require much less explicit interaction, as the users’ needs could be sensed directly from human cognitive system? The current brain-computer interfaces (BCIs) still require their users to actively think a pattern that corresponds to a command, and the computer can then interpret this pattern. On the other hand, this pattern is required to be targeted to a pre-specified command. For example, a classic BCI speller requires its user to attend to a certain letter on the screen and tests all letters presented on the screen to detect and spell the one the user is attending to. I will present a new type of brain-computer interfacing that is based on passive or natural human reactions and directly provides feedback to a computing systems. I will demonstrate applications of such brain-computer interfaces in information retrieval and recommender systems, generative modeling of images, and crowdsourcing. 

Zooming in on Tuukka Ruotsalo 

Which courses do you teach? (BSc and MSc)
I just joined DIKU and will be teaching a couple of lectures on data mining and information visualization in the spring and information retrieval related courses in the future.

Which technology/research/projects/startup are you excited to see the evolution of?
I am most excited about combining machine learning with human-computer interaction, cognitive science and brain research. There is emerging technology now that starts to be deployed in practice, also beyond medical usage. 

What is your favorite sketch from the DIKUrevy?
I joined KU very recently and I’m not familiar with DIKUrevy. It looks like something fun!