Name
Abstract | ||
|---|---|---|
Crowdsourcing offers a new way for mobilizing the collective intelligence and efforts of a large group of people to tackle tasks which cannot be efficiently performed by machines (e.g., video transcription, monitoring road conditions at specific locations) (Doan, Ramakrishnan, and Halevy 2011). Many commercial crowdsourcing platforms are now available. Their business depends on providing satisfactory services to both the crowdsourcers (i.e. task requesters) and the workers involved. From the crowdsourcers' perspective, they expect to receive high quality results for their crowd sourcing tasks in a timely manner. From the workers' perspective, they want to earn as much as possible while committing limited productive effort. As the crowdsourcers are the main source of revenue for crowdsourcing platforms, their requirements tend to take precedence over those of the workers. Many commercial crowdsourcing platforms have implemented some variants of reputation-based mechanisms (Yu et al. 2013a) to gauge the workers' competence based on their past performance, and allow only those with good reputations to access tasks. While this simple reputation-based task delegation method is intuitive and has its own merits, a different, albeit related problem, has not received much attention. As crowd sourcing workers are human beings, they have limited availability and productive capacities to work on tasks delegated to them (Yu et al. 2012). Concentrating requests to workers with good reputations may result in details. In addition, as a small portion of reputable workers become overloaded with task requests while others remain relatively idle, attrition may occur in the worker population, thereby leaving the crowdsourcing platform with a shortage of workers. The potentially conflicting objectives between crowdsourcers and workers can be formalized under the congestion game framework (Monderer and Shapley 1996). In congestion games, the payoff for each player depends on the resources it selects and the number of other players selecting the same resources. For instance, the morning commute to work places by many people can be modeled as a congestion game. The time taken by a traveller on a given day depends on how many others are taking the same route as him. |
Year | Venue | Field |
|---|---|---|
2016 | THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE | Computer science,Crowdsourcing,China,Active living,Knowledge management,Artificial intelligence,Computer Science and Engineering,Library science,Excellence,Machine learning,Beijing |
DocType | Citations | PageRank |
Conference | 2 | 0.35 |
References | Authors | |
10 | 6 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Han Yu | 1 | 639 | 48.71 |
| Chunyan Miao | 2 | 2307 | 195.72 |
| Zhiqi Shen | 3 | 1148 | 82.57 |
| Jun Lin | 4 | 2 | 0.35 |
| Cyril Leung | 5 | 899 | 62.23 |
| Qiang Yang | 6 | 17039 | 875.69 |