Schedule & Papers

The 2nd Workshop on Social Computing and User Generated Content

9:00am-10:00am: Invited Talk

10:00am-10:30am: Coffee break

10:30am-11:30am

11:30am-12:30am: Invited Talk

12:30 - 2:30: Lunch break

2:30 3:30

3:30-4:30: Invited Talk

References

Panos Ipeirotis

Title: Crowdsourcing: Quality Management and Scalability

Abstract

I will discuss the use of crowdsourcing for building machine learning models quickly and under budget constraints, with a focus on the case where humans are noisy and the of "labels" provided by humans for data items are imperfect. I will present strategies of managing quality in a crowdsourcing environment, showing in parallel how to integrate data acquisition with the process of learning machine learning models. I illustrate the results using real- life applications drawn from the field of online advertising. Time permitting, I will also discuss our latest results showing that mice and Mechanical Turk workers are not that different after all.

Bio

Panos Ipeirotis is an Associate Professor at the Department of Information, Operations, and Management Sciences at the Stern School of Business of New York University. His recent research interests focus on crowdsourcing and on mining user-generated content on the Internet. He received his Ph.D. in Computer Science from Columbia University in 2004. He has received three "Best Paper" awards (IEEE ICDE 2005, ACM SIGMOD 2006, WWW 2011), two "Best Paper Runner Up" awards (JCDL 2002, ACM KDD 2008), and is also a recipient of a CAREER award from the National Science Foundation. He also maintains the blog "A Computer Scientist in a Business School" where he blogs about crowdsourcing, user-generated content, and other random facts, and his blogging activity seems to generate more interest and recognition than any of the other activities mentioned in this bio.

Radu Jurca

Title: Peer-driven Incentive Mechanisms

Abstract

Crowdsourcing is an important component of the internet, with profound implications on the future of producing and consuming information. The commercial potential of harnessing the wisdom of the crowds is self-evident; unfortunately, effective mechanisms for quality control (i.e., spam filtering, encourage effort and truthfulness) are less understood. The economic theory offers a framework for designing explicit incentives (monetary or in-kind) able to encourage honest participation in some types of crowdsourcing applications. In this talk, I will survey a family of such incentive mechanisms that are "peer-driven", in the sense that truthfulness is measured against the collective information provided by a user's peers.

Bio

Radu Jurca obtained the Ph.D. degree in Computer Science from Ecole Polytechnique Federale de Lausanne (EPFL) in 2007. His thesis investigates mechanisms for rewarding truthful feedback in online systems, and was awarded the IFAAMAS Victor Lesser Distinguished Dissertation Award (2007) and the EPFL's Best PhD Thesis Award (2008). Radu's research interests focus around the design of feedback and reputation mechanisms in social networks, crowdsourcing applications and other online systems where the information shared by individual participants cannot be verified by a trusted third party. Radu is currently working for Google in Zurich.

Craig Boutilier

Title: Social Choice and Preference Models: New Approaches and Challenges for Group Decision Making

Abstract

Social choice has been the subject of intense investigation within computer science, AI, and operations research, in part because of the ease with which preference data from user populations can now be elicited, assessed, or estimated in online settings. In many domains, the preferences of a group of individuals must be aggregated to form a single consensus recommendation, placing us squarely in the realm of social choice.

The application of social choice and voting schemes to domains like web search, product recommendation, and social networks places new emphasis on issues such as: articulating suitable decision criteria; approximation; incremental preference elicitation; learning methods for population preferences; and more nuanced analysis of manipulation.

In this talk, I'll provide an overview of some of these challenges and outline some of our recent work tackling of them, including: learning probabilistic models of population preferences from choice data; robust optimization (winner determination) with incomplete user preferences; incremental preference elicitation for group decision making; and new analyses of manipulation. I'll also outline challenges and opportunities for exploiting social networks in assessing user preferences.

Bio

Craig Boutilier is a Professor of Computer Science at the University of Toronto. He received his Ph.D from Toronto in 1992, and joined the faculty of University of British Columbia in 1991 (where he remains an Adjunct Professor). He returned to Toronto in 1999, and served as Chair of the Department of Computer Science from 2004-2010. Boutilier has held visiting positions at Stanford, Brown, Carnegie Mellon and Paris-Dauphine, and served on the Technical Advisory Board of CombineNet for nine years.

Boutilier's has published over 180 refereed articles covering topics ranging from knowledge representation, belief revision, default reasoning, and philosophical logic, to probabilistic reasoning, decision making under uncertainty, multiagent systems, and machine learning. His current research efforts focus on various aspects of decision making under uncertainty: preference elicitation, mechanism design, game theory and multiagent decision processes, economic models, social choice, computational advertising, Markov decision processes and reinforcement learning. Boutilier served as Program Chair for both UAI-2000 and IJCAI-09, and is currently Associate Editor-in-Chief of the Journal of Artificial Intelligence Research (JAIR). He is also a Fellow of the Association for the Advancement of Artificial Intelligence (AAAI).