Paper Info
Title
Novel Mechanisms For Online Crowdsourcing With Unreliable, Strategic Agents
Abstract
Motivated by current day crowdsourcing platforms and emergence of online labor markets, this work addresses the problem of task allocation and payment decisions when unreliable and strategic workers arrive over time to work on tasks which must be completed within a deadline. We consider the following scenario: a requester has a set of tasks that must be completed before a deadline; agents (aka crowd workers) arrive over time and it is required to make sequential decisions regarding task allocation and pricing. Agents may have different costs for providing service and these costs are private information of the agents. We assume that agents are not strategic about their arrival times but could be strategic about their costs of service. In addition, agents could be unreliable in the sense of not being able to complete the assigned tasks within the allocated time; these tasks must then be reallocated to other agents to ensure ontime completion of the set of tasks by the deadline. For this setting, we propose two mechanisms: a DPM (Dynamic Price Mechanism) and an ABM (Auction Based Mechanism). Both mechanisms are dominant strategy incentive compatible, budget feasible, and also satisfy ex-post individual rationality for agents who complete the allocated tasks. These mechanisms can be implemented in current day crowdsourcing platforms with minimal changes to the current interaction model.
Year
Venue
Field
2015
PROCEEDINGS OF THE TWENTY-NINTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE
Computer science,Price mechanism,Crowdsourcing,Strategic dominance,Real-time computing,Common value auction,Artificial intelligence,Private information retrieval,Incentive compatibility,Operations research,Mechanism design,Payment,Machine learning
DocType
Citations 
PageRank 
Conference
4
0.43
References 
Authors
5
4
Name
Order
Citations
PageRank
Praphul Chandra1115.06
Yadati Narahari217311.50
Debmalya Mandal3255.59
Prasenjit Dey4262.10