BKT-POMDP: Fast Action Selection for User Skill Modelling over Tasks with Multiple Skills.

Nicole Salomons Emir Akdere Brian Scassellati

Published in: IJCAI (2021)

Keyphrases

real robot
reinforcement learning
action selection
robot soccer
mobile robot
state space
temporal difference
basal ganglia
skills needed
human robot
human users
end users
reinforcement learning algorithms
user profiles
learning algorithm
user model
user interface
multiple users
multiple tasks
user interaction