2021/02/12 by Preeti Ramaraj, Ramaraj, Preeti, Jr. Ortiz +3
Computer Science · #Online Learning and Analytics #Intelligent Tutoring Systems and Adaptive Learning #Reinforcement Learning in Robotics
paper · pdf · doi:10.48550/arxiv.2102.06755
Interactive Task Learning (ITL) is an emerging research agenda that studies\nthe design of complex intelligent robots that can acquire new knowledge through\nnatural human teacher-robot learner interactions. ITL methods are particularly\nuseful for designing intelligent robots whose behavior can be adapted by humans\ncollaborating with them. Various research communities are contributing methods\nfor ITL and a large subset of this research is \robot-centered with a\nfocus on developing algorithms that can learn online, quickly. This paper\nstudies the ITL problem from a \human-centered perspective to provide\nguidance for robot design so that human teachers can naturally teach ITL\nrobots. In this paper, we present 1) a qualitative bidirectional analysis of an\ninteractive teaching study (N=10) through which we characterize various aspects\nof actions intended and executed by human teachers when teaching a robot; 2) an\nin-depth discussion of the teaching approach employed by two participants to\nunderstand the need for personal adaptation to individual teaching styles; and\n3) requirements for ITL robot design based on our analyses and informed by a\ncomputational theory of collaborative interactions, SharedPlans.\n