7月9日：A Structural Review on Neural Networks Based Intelligent Systems

主讲人：William Guo 教授

主持人：周旭辉 教授

开始时间：2016-7-9（周六）上午9:00

讲座地址：资环楼435会议室

主办单位：生态与环境科学学院 科技处

报告人简介：

Dr William Guo is a professor in applied computation and mathematics at Central Queensland University Australia. He received a PhD from The University of Western Australia in 1999, Master of Science in 1991 and Bachelor of Engineering in 1982 in China. His research interests include computational intelligence, data and image processing, modelling and simulation, and geophysics. He has published about 100 papers in international journals, conference proceedings, and edited books, and co-edited two special issues in international journal “Mathematical Problem in Engineering”. He has supervised multiple PhD students and served as a keynote speaker at many international conferences and regional events. He has abundant experience in leadership and academic governance through his services as Dean/Deputy Dean of School, and Members of University Academic Board, Education Committee, and Academic Promotion Committee. He is a member of IEEE, ACM, ACS, and Australian Mathematics Society (AUSTMS).

内容摘要：
Artificial neural networks have been widely used in various intelligent systems either alone or cooperatively with other means for forecasting, classification, decision making, pattern recognition, data analysis, and other purposes in many different disciplines for the last three decades. There have been several reviews that concentrated on some certain aspects of neural networks related systems over the years. This review will summarize the frameworks of neural networks based intelligent systems (NNBIS) in application domain for the past 20 years. It uses a workflow-based logical approach to categorise various NNBISs. This review focuses on applications of, rather than theoretical analysis on and comparisons among, different NNBISs.

In this review, NNBISs are broadly classified into Traditional, Sequential, Concurrent, and Incorporative Structures. Various sub-models of each structure are presented and illustrated using publications from many different disciplines, including applications in ecological, agricultural, and environmental sciences. Each structure is also assessed on its nature for popular applications. Hybrid systems perform not necessarily better than single technique-based models. The choice or design of NNBIS is more problem-oriented, i.e., horses for courses!