Towards Practical Crowdsourcing: Issues and State-of-the-arts

PhD Qualifying Examination


Title: "Towards Practical Crowdsourcing: Issues and State-of-the-arts"

by

Mr. Longfei SHANGGUAN


Abstract:

The popularity and promise of crowdsourcing techniques, as well as the 
success of many crowdsourcing applications on the web, have attracted 
great attention from diverse research communities, including Machine 
Learning, Information Retrieval, Data Mining, Database and Networking, 
etc. Despite the convenience and effectiveness, crowdsourcing also brings 
new challenges, such as making repeated decisions about prices to tasks, 
workers to filter out, and problems to assign. These issues lead to a 
wealth of research works from both theoretical and systematical 
perspectives. In this survey, we conducted a thorough overview of 
crowdsourcing techniques and research issues. Based on different research 
scopes, we classify the research issues overall into three categories: 1): 
Forming the crowd: incentive mechanism design; 2) Evaluating the crowd: 
worker quality control; 3): Optimizing the crowd: task allocation 
optimization. For each category, we investigated a large body of recent 
works with an emphasis on the rationale behind them, and classified them 
into subcategories according to different principles of the methods.


Date:			Tuesday, 28 October 2014

Time:                  	10:00am - 12:00noon

Venue:                  Room 3501
                        Lifts 25/26

Committee Members:	Dr. Ke Yi (Supervisor)
 			Prof. Gary Chan (Chairperson)
 			Dr. Pan Hui
 			Dr. Qiong Luo


**** ALL are Welcome ****