Behavioural Cloning

Behavioural cloning is a method by which human subcognitive skills can be captured and reproduced in a computer program. As the human subject performs the skill, his or her actions are recorded along with the situation that gave rise to the action. A log of these records is used as input to a learning program. The learning program outputs a set of rules that reproduce the skilled behaviour. This method can be used to construct automatic control systems for complex tasks for which classical control theory is inadequate. It can also be used for training

We have developed a behavioural cloning system for aircraft simulators. This is of interest to airlines and simulator manufacturers. For example, it very expensive to train a pilot to fly a 747. Any system that can be used to either eliminate unsuitable candidates early in training or reduce training time will result in substantial savings and an increase in flight safety.

This method has also be used to build control systems for container cranes, manufacturing lines. Our current research is on using behavioural cloning to teach a robot to traverse rough terrain.

Selected Publications

Sammut, C., Hurst, S., Kedzier, D. and Michie, D. (1992). Learning to Fly. In D. Sleeman (Ed.), Proceedings of the Ninth International Conference on Machine Learning, Aberdeen: Morgan Kaufmann, pp. 385-393.

Sammut, C. (1996). Automatic construction of reactive control systems using symbolic machine learning. Knowledge Engineering Review. 11(1): 27-42.

Sammut, C. (1997). A multistrategy approach to learning control from human skills. In D. Brandt (Ed.), Automated Systems Based on Human Skill, (pp. 31-34). Kranjska Gora, Slovenia: IFAC.

Bratko, I., Urbancic, T., and Sammut, C. (1998). Behavioural Cloning of Control Skill. In R. S. Michalski, I. Bratko, and M. Kubat (Eds.), Machine Learning and Data Mining. Wiley, pp. 335-351.

Bain, M. and Sammut, C. (1999). A framework for behavioural cloning. In S. Muggleton, K. Furakawa, and D. Michie (Eds.), Machine Intelligence 15. Oxford University Press.

Isaac, A. and Sammut, C. (2003) Goal-directed Learning to Fly. In T. Fawcett and N. Mishra (Eds). Proceedings of the Twentieth International Conference on Machine Learning. Washington, D.C. pp. 258-265.

Sammut, C., W. Kadous and R. Sheh (2007). Learning to Drive Over Rough Terrain. International Symposium on Skill Science. K. Furukawa. Tokyo.

Sheh, R., Hengst, B., and Sammut, C. (2011). Behavioural cloning for driving robots over rough terrain. In International Conference on Intelligent Robots and Systems (IROS), pages 732–737, San Francisco, CA.