Knowledge Discovery From Social and Information Networks
Descripción general del curso
The first part of the course will focus on understanding machine learning algorithms and identifying challenging problems on the Web, learning how to apply machine learning algorithms to these problems, and how to use the existing tools and design new ones. Examples of topics include: supervised learning techniques, e.g., text classification, kernel methods and Support Vector Machines, Bayesian learning, Artificial Neural Networks and Deep Learning techniques, as well as semi-supervised learning techniques.
The second part of the course will focus on modeling social networks. Examples of topics include: what are networks and why do we study them; describing and measuring networks (e.g., centrality, degrees, diameters); community (e.g., clustering, community structure); opinion mining, coordination and cooperation. Each lecture will include a guided, hands-on exercise for students using publicly-available machine learning and data mining tools on large document collections obtained from well-known digital library portals and social media sites.
Prerequisites: Basic knowledge on probability and statistics, data structures, programming, and algorithms. Background in machine learning or social and information network analysis is not required.
Información del curso
7 de Junio al 22 de Junio
Cornelia Caragea, University of North Texas
2:00 p.m. a 6:00 p.m.
Curso del perfil PNIF
Estudiantes otras maestrías:
Institución: University of North Texas
Hoja de vida
The design of accurate approaches for automatic information extraction and scienti c data analysis, which can foster knowledge discovery and organization; the design of supervised and semi-supervised learning approaches to handle today's very large collections of data from many data miningapplications; and the design of privacy analysis technologies that can help users to maintain control of their privacy in online environments, while allowing users to increase their online social capital without leading to overexposure. This research has resulted in peer-reviewed publications in top venues, e.g., ACM Transactions on the Web (TWeb 2015); the American Association for Arti cial Intelligence (AAAI 2016, 2014); International World Wide Web (WWW 2015, 2013); and Empirical Methods in Natural Language Processing (EMNLP 2015, 2014); as well as invited talk at top universities,e.g., University of Texas at Austin, and University of Michigan, Ann Arbor, and at events such as the Big Scholarly Data co-located with the Microsoft Research Faculty Summit in Redmond.
Two papers that Cornelia co-authores were nominated for best paper awards and one of them received the \Most Innovative Application of AI" award at the Conference on Innovative Applications of Arti cial Intelligence (IAAI 2014). Another part of my work performed with two high-school students received the 3rd place in Computer Science at the Fort Worth Regional Science and Engineering Fair 2015.
Cornelia collaborates with many universities and industry labs, including Pennsylvania State University, University of Michigan, University of the And es, University of Bucharest, Yahoo Labs, Microsoft Research, and A* STAR Infocomm Research, Singapore.Her research is currently supported by several grants from the National Science Foundation.