Tutorial Sessions/Invited Talks

Use of Social Media platforms is growing at an alarming rate. Over the past years, we have seen a continuous growth in the amount of social media data being generated as well as increasing sophistication and complexity of social media platforms. This has spawned a vibrant community, both in academia and industry, looking to make sense of what is going on. While the amount of attention in this space is growing, the analytic techniques are still in their infancy and the research questions are still quite narrow. In fact, most work is focused either on the social networks or on the content and the individual users. I will in this talk try to expand beyond the majority of current research and discuss how we can make use of both the content and the network to better understand and mine social media. Specifically, I will here talk about how we can automatically generate psychographic profiles of users based on analyzing their generated content, and then how we can use these models to better explain their observed retweeting behaviors.

Sofus A. Macskassy is a Sr. Computer Scientist at the Information Sciences Institute and an Adjunct Professor of Computer Science at the University of Southern California. He was previously the Director of
Fetch Labs at Fetch Technologies. His work spans a wide area of research and domains, including social media analytics, where he is extracting and analyzing user profiles and behaviors from Twitter. His main research areas include statistical relational learning, information filtering, data mining and social network analysis. He is published in top venues in AI, ML and DM and has served on numerous organizing and programming committees. He serves on the editorial board of the Machine Learning Journal, the premier machine learning journal in the world. He is also the primary developer and maintainer of the open-source Network Learning Toolkit for Statistical Relational Learning.

Note: The planned talk 'Crowdsourcing, Human Computation, and Collective Intelligence' will be held as a WORLDCOMP keynote on Monday, July 16 in the morning.

The introduction of crowdsourcing platforms such as Amazon Mechanical Turk has made it possible to harness cheap, albeit sometimes unreliable, human labor in a range of data annotation tasks. The most common approach for improving the quality of annotations is to seek multiple annotations for each item and use majority vote to assign a final output. However, each annotation costs money, and we want to make sure we spend that money in the most effective ways possible. For example, if you are given a fixed budget, what's the best way to allocate data annotation work to crowdsourced labor to achieve the most accurate results that you can? This talk will discuss recent results on how we can improve the labor efficiency of crowdsourced data annotation tasks.

Haym Hirsh is Professor of Computer Science at Rutgers University. His research has focused on foundations and applications of machine learning, data mining, information retrieval, and artificial intelligence, especially targeting questions that integrally involve both people and computing. Most recently these interests have turned to crowdsourcing, human computation, and collective intelligence. From 2006-2010 he served as Director of the Division of Information and Intelligent Systems at the National Science Foundation, and from 2010-2011 he was a Visiting Scholar at MIT's Center for Collective Intelligence. Haym received his BS from the Mathematics and Computer Science Departments at UCLA and his MS and PhD from the Computer Science Department at Stanford University.

Data mining has been progressively expanding from academia and research to commercial and governmental organizational environments. Wide-ranging data mining techniques and analytics are finding applicability in organizational arena. Numerous organizations have been considerably benefiting from implementation of data mining solutions and analytics. However, despite significant advances in data mining techniques, and their efficient implementation methods, majority of organizations are still reluctant to adopt them. What limits adoption of data mining techniques in real-world organizational environments? Which data mining techniques are currently at the forefront of acceptance by diverse organizations? What are the promising trends? These are the pertinent questions we shall explore—after a concise overview of the status quo.

Dr. Peter Geczy is a chief scientist at The National Institute of Advanced Industrial Science and Technology (AIST). He also held positions at The Institute of Physical and Chemical Research (RIKEN) and The Research Center for Future Technologies. His interdisciplinary scientific interests encompass domains of data and web mining, human interactions and behavior, social intelligence technologies, privacy, information systems, knowledge management and engineering, artificial intelligence, and adaptable systems. His recent research focus also extends to the spheres of service science, engineering, management, and computing. He received several awards in recognition of his accomplishments. Dr. Geczy has been serving on various professional boards and committees, and has been a distinguished speaker in academia and industry.

Robert Stahlbock
General Conference Chair

E-mail: conference-chair@dmin-2012.com

Robert Stahlbock. Gary M. Weiss

Programme Co-Chairs

E-mail: programme-chair@dmin-2012.com

This website is hosted by the Lancaster Centre for Forecasting at the Department of Management Science at Lancaster University Management School.