当前位置:首页  教学科研

学术讲座【Hyperrelations–lattice-based learning machine, contextual probability and similarity kernels】

时间:2013-07-11浏览:691设置

时间:2013年7月18日(星期四)上午9:00

地点:仓山校区成功楼603室

主讲:University of Ulster,UK  王晖教授

主办:数学与计算机科学学院

专家简介:Hui Wang is Professor of Computer Science at the University of Ulster, UK. He receives his BSc and MSc degrees in computer science from Jilin University and PhD in artificial intelligence from the University of Ulster. His research interests include theoretical aspects of learning and reasoning and applications in text, video, time series and spectral data analysis. His research in these areas has been funded by local, national and international funding bodies, and has resulted in over 170 archival publications. He was a Tang Aoqing Scholar at Jilin University and was a visiting professor at Tsinghua University. He is an associate editor for IEEE Transactions on Cybernetics, International Journal of Computational Intelligent Systems and International Journal of Machine Learning and Cybernetics.

报告摘要:Hyperrelations are a generalisation of relations from value-based to set-based. Hyperrelations can be used as knowledge representation, and as the means of reasoning via the notions of contextual probability and kernel similarity. More specifically, hyperrelations can be used for approximation based learning which allows incremental learning; for density based clustering which allows automatic determination of the cluster number; for contextual probability estimation which is non-parametric, and does not require conditional independence of attributes; and for designing similarity kernels that lend themselves for use in kernel machines. Hyperrelations and other related concepts/operations can also be used in text and video analysis.

返回原图
/