Analysis of crowdsourcing systems
Crowdsourcing is envisioned as a paradigm that can potentially change the future job market substantially, gaining a significant footprint in the global economy. Indeed, it provides highly flexible workforce and mitigates the effects of possible scarcity of experts in a given geographical area, by allowing the employers to get in touch with the most qualified workers worldwide. Furthermore, it can generate new opportunities for increasing both productivity and social mobility in regions where investments are small and the economy is stagnant. However the current crowdsourcing model hardly supports the complexity, creativity and skills that are needed to solve most of the relevant professional problems, today. Significant improvements are needed to make crowd work a viable alternative to traditional labor systems. Our activity stems from these considerations. Through a system theoretic approach that addresses several difficult research challenges, it aims at the design of new generation crowdsourcing platforms, which can support of highly valuable work. The ultimate goal of our activity is to devise a set of strategies/algorithms/architectures that make crowdsourcing systems, attractive for the solution of highly valuable problems. Examples of technical questions that we wish to answer are: how can my problem be optimally partitioned in tasks? How task dependencies can be modeled? How can tasks be efficiently assigned to workers? How should my workers be organized? Which is an efficient pricing strategy to motivate workers? How can I discourage malicious workers' behaviors? How can I design an efficient system of incentives/rewards (economic and not economic) and sanctions? Which is the best way to process workers' output and reach a decision? How can I estimate the reliability degree of my solution? How can I effectively process jobs with real-time constraints? How can I ensure worker's privacy and anonymity?
- PE7_8 Networks (communication networks, networks of sensors, robots...)
- PE7_9 Man-machine-interfaces
- LEONARDI EMILIO - Manager
- AJMONE MARSAN MARCO GIUSEPPE