My research interests span distributed systems, networking and databases, including parallel computing, and the recent work focuses on large-scale graph data processing (e.g. queries on connectivity and machine learning). My current focus is on auto-tuning of data processing/analytics framework to deal with complex parameter space using machine learning approaches. Various machine learning approaches (e.g. Deep Neural Networks, DNN) are emerging for data analysis. Powerful accelerators demonstrate superb performance in training DNNs compared to CPUs. Users are in great need of tools for tuning such data processing frameworks. The multidimensional design space for optimising applications is huge. Techniques for load balancing, job scheduling and adaptive processors require run-time optimisations that depend on the dynamics of computation resources, input data and running algorithms. Computations running massive data vary and not all computation can be processed as input data in matrix formation. Emerging graph analytics will provide powerful insights. Optimisation tasks are complex, and the run-time system gains new responsibility for appropriate workload configuration.
Another project I focus is Digital Epidemiology, where understanding of human interactions in rural societies of developing countries by applying wireless technology and advanced modelling of infectious disease spread. The results will provide unique insights into key epidemic and statistical issues. Ultimately, this research could contribute to the prevention and control of infectious diseases by quarantining coalescing hubs and contributing to early-outbreak alerting systems. The work is tightly related to recent GCRF support.