Abstract:
We develop an on-line method for learning robot behaviours that requires only a small number of trials, making it practical for use on a real robot. The method uses a symbolic planner, including qualitative reasoning, to construct an approximate solution to a control problem.
This approximate solution provides constraints for numerical optimisation, which is used to refine the qualitative plan into an operational policy. The method is demonstrated on a multi-tracked robot intended for urban search and rescue.
Bio:
Claude Sammut is a Professor of Computer Science and Engineering at the University of New South Wales, Head of the Artificial Intelligence Research Group and Deputy Director of the iCinema Centre for Interactive Cinema Research. Previously, he was a program manager for the Smart Internet Technology Cooperative Research Centre, the UNSW node Director of the ARC Centre of Excellence for Autonomous Systems and a member of the joint ARC/NH&MRC project on Thinking Systems.
His early work on relational learning helped to the lay the foundations for the field of Inductive Logic Programming (ILP). With Donald Michie, he also did pioneering work in Behavioural Cloning. His current interests include Conversational Agents and Robotics. He was the leader of the UNSW teams that won RoboCup four-legged robot competitions in 2000, 2001 and 2003 and the CAS team that won the award for best autonomous robot at RoboCup Rescue 2009 – 2011. In 2014, the UNSW team, rUNSWift became champions of the Standard Platform League.
Claude Sammut has been a member of the editorial boards of the Journal of Machine Learning Research, the Machine Learning Jourmal and New Generation Computing. He was the program and general chair of the 2002 International Conference on Machine Learning and the general chair of ICML 2007. He was a member of the executive committee of the RoboCup Federation from 2003 to 2009 and in 2012 was elected to the board of trustees of the RoboCup Federation and is co-editor-in-chief of Springer’s Encyclopedia of Machine Learning.