فارسی
Sunday 20 October 2019

Cardinal Theme

An open model for question answering systems based on Crowdsourcing

Today people use the Internet to find the answer to their questions. They mostly rather ask other users on Community Question Answering (CQA) sites for an answer than just searching the web. However, as Social Media becomes more popular, users tend to ask their questions on these networks, and ignore the benefits CQA sites offer. On the other hand, automatic Question Answering (QA) systems are unable to comprehend questions including images and implementing necessary algorithms for such systems is expensive. In this paper, we propose QA process based on Crowd sourcing, which runs on a QA open system. The system benefits from Crowd sourcing advantages, besides automation techniques. The model is operational and we have demonstrated that questions could be received from different heterogeneous sources, if the suitable procedures are used, and that the answer is obtained from the crowd in the proposed process based on Crowd sourcing. Moreover, the first Iranian crowd sourcing platform for complicated tasks is implemented, which could be used as a basis for future research.


Mohammad Javad Kargar
Abbas Oveissi


IEEE