Research Challenges to Semi-Automatically Enhance Quality in Distributed Open Archives

Abstract:

The "Open Archives Distributed" project proceeds at University of Oldenburg and at Virginia Tech, funded by DFG and NSF, respectively. It concerns managing both physics information and electronic theses/dissertations in Germany and USA, building upon the Open Archives Initiative. In this talk the focus is on recent work at Virginia Tech, especially the "Reengineering PhysNet in the uPortal Framework" thesis work of Ye Zhou, which offers a number of new services: PACS browsing, PACS automatic classification (using SVM together with extensive training), Physics employment service, XML schema-driven web-based metadata generation, and user account management. We seek advice on future efforts, such as: usability testing with physicists, browsing, filtering, PACS-based recommending, better crawling to accurately identify and collect physics department information, high-performance searching, and multi-lingual support.

DOWNLOAD PRESENTATION (pdf)