作者: Christoph Lofi , Kinda El Maarry
DOI: 10.1109/CBI.2014.16
关键词:
摘要: Crowd sourcing has shown to be a powerful technique for over-coming many challenges in data and information processing where current state-of-the-art algorithms are still struggling. This is especially true workflows that transparently combine algorithmic heuristics dynamically crowd sourced tasks performed by human workers, which promise solve even more complex effectively efficiently. But still, such hybrid can difficult approach, they often designed an ad-hoc fashion. Therefore, this paper, we extensively investigate as described the literature, abstract generic design patterns, codify commonly recurring their best-practice solutions. Each pattern discussed with special focus on its requirements, constraints, effects overall workflow. We illustrate practicality of these patterns providing real-world application examples or have been applied. Furthermore, show-case how individual extended combined support workflows. Our provide extensive overview workflows' space, allow efficient modeling, analysis, documentation