Data Leak Aware Crowdsourcing in Social Network

TL;DR

This paper introduces Friendlysourcing, a method for forming teams in social networks for crowdsourcing tasks while preventing partial solution leaks to competitors, using clustering and Markov-chain algorithms.

Contribution

It presents a novel clustering approach and a Markov-chain algorithm to identify collaborative and competitive groups in social networks.

Findings

Effective detection of competitive groups in social networks.

Successful prevention of partial solution disclosure.

Improved team formation for crowdsourcing tasks.

Abstract

Harnessing human computation for solving complex problems call spawns the issue of finding the unknown competitive group of solvers. In this paper, we propose an approach called Friendlysourcing to build up teams from social network answering a business call, all the while avoiding partial solution disclosure to competitive groups. The contributions of this paper include (i) a clustering based approach for discovering collaborative and competitive team in social network (ii) a Markov-chain based algorithm for discovering implicit interactions in the social network.