Skip to main content
Robot from the CTDS labratory
Research_

The science of decision-making

How to make the right decision under pressure
Having to make important decisions causes headaches for many people. Our team studies the science of decision-making, creating research that will enable everyone from doctors to governments make better decisions.

We study the science of decision-making under uncertainty. Given probabilistic predictions of future events, how do we choose a sequence of decisions to maximise a particular objective? For example, given a dataset with diseases and outcomes of treatments, what treatments should be selected to improve the patient’s recovery? Our research is currently focused on multi-arm bandit settings, Bayesian optimisation and reinforcement learning. 

Techniques

  • Partially observable Markov decision process
  • Reinforcement learning
  • Bayesian optimisation
  • Experimental design
  • Planning

Current projects

  • Machine learned defence to adversarial RF systems via system identification and strategy optimisation
  • Bayesian risk prediction instruments
  • Space-time and demographics crime prediction
  • Academic capability mapping
  • Bayesian optimisation for GLM applied to police patrolling
  • Path planning in dynamic environments – South Korean Agency for Defense Development
  • Galactose metabolomics

Completed projects

  • A whole-business data strategy for Sydney Water
  • Dynamic adipocyte metabolomics
  • Transient candidate vetting using convolutional neural nets