CSR---PDOS: Information Pedigree

Regents Of The University Of Michigan - Ann Arbor, Ann Arbor MI

Investigators

Abstract

As the scope of available information grows, it is increasingly difficult to find what one needs when it is needed. The growth of search as an industry is testament to this problem, but existing tools are incomplete solutions. They index content, but not context. They capture only static, syntactic relationships, not dynamic, semantic ones. Perhaps most importantly, they have access only to information stored either on a user's local machine or publicly-accessible distributed stores---other sources are neither indexed nor searchable. Information Pedigree solves this problem by capturing context and semantic relationships between information items, collecting disparate relationship meta-data to support searches, and providing for availability of relevant items created on machines controlled by individual users. Information Pedigree achieves this goal by instrumenting the flow of information between a user's applications and services, as well as the flow of information between different users. By observing these transactions, we form a contextual and semantic index over data that is used, simplifying its retrieval for later use. Information Pedigree will make it easier for people to find and share data, without placing additional burdens on them or their IT support staff. This greatly reduces time wasted in managing information, allowing information workers to be more productive. The results from this project will be widely disseminated in the research literature, and a prototype of the system will be made available to other researchers as well as users. This project will train several graduate students, and influence the curricula of several graduate and undergraduate courses.