报告人：高俊斌教授（University of Sydney）
报告题目：Matrix Neural Networks
报告摘要：Traditional neural networks assume vectorial inputs as the network is arranged as layers of single line of computing units called neurons. This special structure requires the non-vectorial inputs such as matrices to be converted into vectors. This process can be problematic. Firstly, the spatial information among elements of the data may be lost during vectorisation. Secondly, the solution space becomes very large which demands very special treatments to the network parameters and high computational cost. To address these issues, we propose matrix neural networks (MatNet), which takes matrices directly as inputs. Each neuron senses summarised information through bilinear mapping from lower layer units in exactly the same way as the classic feed forward neural networks. Under this structure, back prorogation and gradient descent combination can be utilised to obtain network parameters efficiently. Furthermore, it can be conveniently extended for multimodal inputs. We apply MatNet to MNIST handwritten digits classi_cation and image super resolution tasks to show its e_ectiveness. Without too much tweaking MatNet achieves comparable performance as the state-of-the-art methods in both tasks with considerably reduced complexity.
报告人简介：Professor Junbin Gao is an International Standard expert in Machine Learning. His current research interests cover Statistical Machine Learning, Bayesian Inference, Computer Vision and Image Analysis, Data Mining and Big Data, Numerical Optimization, and Visualization. His research on Dimensionality Reduction was featured in an article published on 4 December 2012 by The Australian newspaper. Professor Junbin Gao is an active scholar in major smart computation on the world stage. He has been serving as an assessor of international standard for ARC since 2004. He is an associate editor for Journal of Probability and Statistics and Applied Computational Intelligence and Soft Computing. He serves as local chair or program member for more 50 international conferences such as CVPR, PAKDD, IEEE SMC etc.
报告时间：2017年1月3日上午9:00-10:30
报告地点：科技楼602
