Cristian Rojas

Cristian Rojas was born in Manchester, England, got an MS in Electronics Engineering from UTFSM, Valparaíso, Chile, in 2004, and a PhD in Electrical Engineering at The University of Newcastle, Australia, in 2008.

Powerful computers, new sensors and highly efficient communication technology have led to the availability of large amounts of data and opened the door to the design of engineering systems that can learn by themselves. As an example, several years ago, Google’s self-taught AlphaGo computer defeated the world champion of Go. The development of self-learning systems is happening on many fronts, including factories, smart buildings and autonomous vehicles. In most of these applications, physical systems or humans interact with a computer, which has a mathematical model of its environment. Complicating factors are that physical systems and humans have dynamics, that is, their behaviour depends on what has happened before, and that the data available from them is not always insightful or useful for learning.
Cristian’s research lies at the intersection of several fields that deal with the most efficient use of data for the design of such self-learning systems, including control theory, system identification, statistics and machine learning. His goal is to analyse and design methods for efficiently learning and controlling dynamical systems from data and prior knowledge, such as physical insight or existing models for such systems.