Controlling a complex dynamic system, such as a plane or a crane, usually requires a skilled operator. Such a control skill is typically hard to reconstruct through introspection. Therefore an attractive approach to the reconstruction of control skill involves machine learning from operators' control traces, also known as behavioural cloning. In the most common approach to behavioural cloning, a controller is induced in the form of a rule set or a decision or regression tree that maps system states to actions. Unfortunately, induced controllers usually suffer from lack of robustness and lack typical elements of human control strategies, such as subgoals and substages of the control plan. In this paper we present a new approach to behavioural cloning which involves the induction of a model of the controlled system and enables the identification of subgoals that the operator is pursuing at various stages of the execution trace. The underlying formal basis for the present approach to behavioural cloning is the theory of LQ controllers. Experimental results show that this approach greatly improves the robustness of the induced controllers and also offers a new way of understanding the operator's subcognitive skill.
[1]
Donald Michie,et al.
Knowledge, Learning and Machine Intelligence
,
1993
.
[2]
Ivan Bratko,et al.
Behavioural Cloning: Phenomena, Results and Problems
,
1995
.
[3]
Claude Sammut,et al.
Learning to Fly
,
1992,
ML.
[4]
Steven J. Bradtke,et al.
Reinforcement Learning Applied to Linear Quadratic Regulation
,
1992,
NIPS.
[5]
Dimitri P. Bertsekas,et al.
Dynamic Programming: Deterministic and Stochastic Models
,
1987
.
[6]
Ivan Bratko,et al.
Reconstructing Human Skill with Machine Learning
,
1994,
ECAI.
[7]
Stefan Schaal,et al.
Assessing the Quality of Learned Local Models
,
1993,
NIPS.