CiNet Monthly Seminar
December 20, 2017
16:00 ~ 17:00
CiNet 1F Conference Room
“Understanding reinforcement learning on a deeper level”
Sang Wan Lee
Department of Bio and Brain Engineering, KAIST
Host : Ben Seymour (PI)
Latest research in reinforcement learning (RL) has demonstrated an ability to succeed in a few arduous tasks. However, fundamental questions still remain as to how the human brain develops an ability to handle a wide variety of tasks and to learn from only few observations. This talk introduces our research team’s twofold approach to advancing the understanding of human RL, by juxtaposing wisdoms from neuroscience and AI.
The first line of research (AI to brain) examines neural correlates of reinforcement learning, causal inference, and prefrontal meta-control. I will summarize our recent findings to support the view that the brain implements multiple distinctive types of learning and inference: model-based and model-free RL, incremental and one-shot inference. This idea forms a basis for the theory that one of the key functions of human prefrontal cortex is to allocate behavioral control to the brain’s subsystems, placing it as the “prefrontal meta-controller”.
The second line of research (Brain to AI) focuses on brain-inspired RL algorithms. I will then discuss an applicability of RL to engineering and neuroscience: AI-human co-evolution engine and neural experimenter. A detailed insight into these issues not only permits advances in AI, but also helps us understand the nature of human intelligence on a deeper level.
Program of Brain and Cognitive Engineering
KAIST Institute for Health, Science, and Technology
KAIST Institute for Artificial Intelligence
Korea Advanced Institute of Science and Technology (KAIST)
PhD, KAIST (2009)
Postdoc, MIT (2010-2011), Caltech (2011-2015)
Assistant professor, KAIST (2015-now)
Della-Martin fellowship (2014),
KIIS Young Investigator Award (2016), ICROS Young Investigator Award (2016),
Google faculty research award in computational neuroscience (2017)
Research interest: computational neuroscience, brain-inspired AI
Sang Wan Lee is currently an assistant professor with the department of bio and brain Engineering at KAIST, and the director of the laboratory for brain and machine intelligence (http://aibrain.kaist.ac.kr). In 2009, he received Ph.D. in Electrical Engineering and Computer Science from KAIST. During 2010-2015, he was a postdoctoral associate at Mcgovern institute for brain research at MIT, followed by a Della Martin postdoctoral scholar in the Computation & Neural Systems and the Behavioral & Social Neuroscience program at Caltech. He was the recipient of the Della-Martin fellowship (2014) and the Google faculty research award for computational neuroscience (2017). His research interests include brain-inspired artificial intelligence and computational neuroscience.
#14 January 10 (Wed.): Neural networks as a window on human cognition
Marcel van Gerven, Donders Institute for Brain, Cognition and Behaviour, Radboud University
#15 January 12 (Fri.): Out of body, out of mind: the bodily self and embodied cognition
Henrik Ehrsson, Department of Neuroscience, Karolinska Institutet