Supporting complex work in crowdsourcing platforms: a view from service-oriented computing

Download files
Access & Terms of Use
open access
Copyright: Xiao, Lu
Today, crowdsourcing is changing the way people work and solve problems - from ”in-house working” to ”public outsourcing”. Many online crowdsourcing platforms allow the requester to advertise their tasks and help crowd workers find their jobs. However, those platforms mainly focus on the micro-task market and do not support the complex work consisting of interdependent, professional tasks. To crowdsource this work, we need a way to model professional workers, define the complex task, and a coordination mechanism to manage them. To this end, we propose a service-oriented crowdsourcing framework in this thesis wherein (1). Each professional crowd worker is modelled as a service that can be self-described, dynamically discovered and assembled into the complex crowd work; (2). A complex crowd task is defined as a schema consisting of a set of units of work and their inter-dependencies, which can be used to get multiple crowd workers involved and guide their work; (3). The whole crowdsourcing lifecycle is divided into two phases: (i). Plan Phase - where the working plan on the advertised complex task is crowdsourced to generate the schema as mentioned earlier detailing the original advertising; and subsequently, this schema is transformed into a web service orchestration specification for the later auto-coordination; (ii). Execution Phase - where the execution of the planning result is crowdsourced to complete the original complex work as advertised through coordinating and interacting with multiple crowd workers, based on coordination and interaction protocol.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Xiao, Lu
Paik, Hye-Young
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
Resource Type
Degree Type
Masters Thesis
UNSW Faculty
download public version.pdf 6.52 MB Adobe Portable Document Format
Related dataset(s)