2018 Projects

Bayesian inference via MCMC and Evolutionary Algorithms

• Improve MCMC via Cooperative Coevolution (CC) (Use G3-PCX CC and give a MCMC framework for constrained real parameter optimisation and inference)

• MCMC with Parallel (Multi-threaded in C++) CC for Expensive Parameter Optimisation

Computer Vision via Deep Learning

• Face Recoginition for Authentication via Multi-task learning (Use Videos and Sound)

Bayesian Deep Learning

• Robust Bayesian Neural Networks for Dynamic Time Series Prediction
• Autonoumous feature ranking and learning for modular pattern classifiation via Bayesian Neural Networks

Applications

• Cyclone Track Prediction
• Reef Modelling
• Solid Earth Evolution
• Mineral Exploration