【学术视频】第六届复杂系统统计物理与数学国际研讨会 | 中科院物理所张潘研究员

更多精彩视频登陆网站

www.koushare.com

图 | 张潘

题   目：Phase transitions and optimal algorithms in the semi-supervised classifications in graphs

报告人：张潘

单   位：中国科学院理论物理研究所

时   间：2020-01-15

地   点：华侨大学

扫码观看精彩报告视频

报告摘要

By analyzing Bayesian inference of generative model for random networks with both relations (edges) and node features (discrete labels), we perform an asymptotically exact analysis of the semi-supervised classfication problems on graph-structured data using the cavity method of statistical physics. We unveil detectability phase transitions which put fundamental limit on ability of classfications for all possible algorithms. Our theory naturally converts to a message passing algorithm which works all the way down to the phase transition in the underlying generative model, and can be translated to a graph convolution neural network algorithm which greatly outperforms existing algorithms including popular graph neural networks in synthetic networks. When applied to real-world datasets, our algorithm achieves comparable performance with the state-of-the art algorithms. Our approach provides benchmark datasets with continuously tunable parameters and optimal results, which can be used to evaluate performance of exiting graph neural networks, and to find and understand their strengths and limitations. In particular, we observe that popular GCNs have sparsity issue and ovefitting issue on large synthetic benchmarks, we also show how to overcome the issues by combining strengths of our approach.

个人简介

张潘，本科毕业于兰州大学理论物理专业；2000年9月-2009年7月在兰州大学完成硕博连读；2009年10月-2015年9月，分别在意大利Politecnico di Torino，法国巴黎E.S.P.C.I.，美国Santa Fe Institute做博士后研究工作；2015年9月-2018年12月任中科院理论物理研究所副研究员；2019年1月至今任中科院理论物理研究所研究员。其研究方向为统计物理与机器学习的交叉领域。目前研究兴趣包括自旋玻璃理论，消息传递算法，随机矩阵，统计推断，以及网络和机器学习中的一些基础问题。

会议简介

会议题目：第六届复杂系统统计物理与数学国际研讨会（SPMCS2020)
会议时间：2020-01-14会议地点：厦门主办方：华侨大学

The 6th International Workshop on Statistical Physics and Mathematics for Complex Systems (SPMCS 2020) held at Huaqiao University, Xiamen, China, January 14-18, 2020. SPMCS has begun in Le Mans (France) in 2008, and then has been held in Wuhan (China) in 2010, Kazan (Russia) in 2012, Yichang (China) in 2014, and Wuhan (China) in 2017. The primary aim of this workshop series is to offer a forum to the researchers of complex systems science for having discussions and exchanging ideas and information about the latest developments in the relevant fields.

—— ——往期精彩回顾—— ——

为满足更多科研工作者的需求，蔻享平台开通了各科研领域的微信交流群。

进群请添加微信：18256943123

小编拉您入群哟！

未经许可，禁止转载本页内容

转载授权请微信联系：18256943123

戳这里，观看精彩视频哟！