I (Kai Krueger) am a post doc in Randy O'Reilly's lab and joined in 2011. Previously I was a grad student at the Gatsby Computational Neuroscience Unit at University College London, supervised by Peter Dayan.
My main research interests lie in modelling of prefrontal cortex and executive control. In particular, I'm interested in the amazing flexibility of human control and its ability to rapidly adapt to novel tasks, either through efficient trial and error learning or through the help of instruction. I am interested in identifying some of the computational principals and requirements that allows the brain to achieve these and see if any of these principals can be applied to artificial learning agents and neural networks.
Krueger K A and Dayan P (2009) Flexible shaping: How learning in small steps helps Cognition, 110(3), 380-294
- Computational Cognitive Neuroscience Conference, 2009
Adaptive state space construction as a basis for improved learning through shaping
- Computational and Systems Neuroscience, 2009
Abstract rule representations in a bilinear model
- Computational and Systems Neuroscience, 2008
Modularization through the prism of shaping
- Neural Information Processing Systems, Workshop talk 2007
Flexible Shaping: How learning in small steps helps
- Computational and Systems Neuroscience, 2007