Toward a Computational Theory of Data Acquisition and Truthing
COLT '01/EuroCOLT '01 Proceedings of the 14th Annual Conference on Computational Learning Theory and and 5th European Conference on Computational Learning Theory
| Hi-index | 0.00 |
Open data collection over the World Wide Web-in which any web user can contribute to large databases of "informal" data -presents several challenges that require novel approaches in human interface design, algorithmic machine learning and collaborative infrastructure. Foremost among these challenges is the need to ensure data integrity and quality, by automatically or semi-automatically identifying unreliable or hostile contributors and rejecting their contributed data. Other, traditional requirements include security, scalability, and the need to support collaboration between separate but related data collection projects. Further, for the open software/data model, the system must permit users to freely browse, navigate and download the contributed data. This paper presents a collaborative architecture for a set of such data collection projects that addresses these challenges and requirements. Key components of this architecture have been implemented and tested as part of the Open Mind Initiative.