Reliable classification by unreliable crowds
- 1 May 2013
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 5558-5562
- https://doi.org/10.1109/icassp.2013.6638727
Abstract
We consider the use of error-control codes and decoding algorithms to perform reliable classification using unreliable and anonymous human crowd workers by adapting coding-theoretic techniques for the specific crowdsourcing application. We develop an ordering principle for the quality of crowds and describe how system performance changes with the quality of the crowd. We demonstrate the effectiveness of the proposed coding scheme using both simulated data and real datasets from Amazon Mechanical Turk, a crowdsourcing microtask platform. Results suggest that good codes may improve the performance of the crowdsourcing task over typical majority-vote approaches.Keywords
This publication has 10 references indexed in Scilit:
- Privacy and Reliability in Crowdsourcing Service DeliveryPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2012
- Budget-optimal crowdsourcing using low-rank matrix approximationsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2011
- Quality management on Amazon Mechanical TurkPublished by Association for Computing Machinery (ACM) ,2010
- Who are the crowdworkers?Published by Association for Computing Machinery (ACM) ,2010
- Are your participants gaming the system?Published by Association for Computing Machinery (ACM) ,2010
- Visual Recognition with Humans in the LoopLecture Notes in Computer Science, 2010
- Cheap and fast---but is it good?Published by Association for Computational Linguistics (ACL) ,2008
- Performance Analysis and Code Design for Minimum Hamming Distance Fusion in Wireless Sensor NetworksIEEE Transactions on Information Theory, 2007
- Distributed fault-tolerant classification in wireless sensor networksIEEE Journal on Selected Areas in Communications, 2005
- Distributed Detection and Data FusionPublished by Springer Science and Business Media LLC ,1997