Prof. Maozhen Li

8:30, Dec. 23, 2015

Hall 208, School of Computer Science and Communication Engineering

Brief Introduction

Maozhen Li is a Professor of High Performance Computing (HPC) in the Department of Electronic and Computer Engineering at Brunel University London, UK. He started his research career in HPC in 1999 at Cardiff University, UK where he worked as a Post-Doctoral Research Associate on the EPSRC funded project in developing a software infrastructure for parallel and distributed problem-solving environments for large-scale molecular dynamics simulations. Recently, he has successfully secured three high performance Xeon Servers from Intel featuring in total 104 CPU cores, 256GB RAM and 5TB hard disk space. Currently, he is a Co-Investigator of the EPSRC funded research project looking at big data analytics for smart grid, which is in collaboration with UK National Grid, Alstom and Intel. He has over 100 scientific publications.

This lecture starts with a brief introduction to MapReduce which has become a major computing model in support of big data analytics. It then looks at how a back-propagation neural network can be parallelized using the MapReduce model in enabling large-scale machine learning. Three designs are considered in the parallelization process. Finally, it evaluates the performance of the parallel neural network from the aspects of both computation efficiency and accuracy.