- From: Thompson, Bryan B. <BRYAN.B.THOMPSON@saic.com>
- Date: Wed, 5 May 2004 07:51:40 -0400
- To: "'public-rdf-dawg@w3.org'" <public-rdf-dawg@w3.org>
- Cc: "'kendall@monkeyfist.com'" <kendall@monkeyfist.com>
BT01 - Intelligence Analysis Use Case Sally is focusing her doctoral studies on the application of Social Network theory to perform an analysis of the Internet Movie Database (IMDB) in an attempt to predict the success of movies and the likelihood of collaboration between directors and actors. This is an interesting problem not only for Hollywood pizzazz, but also because it is similar to a counter-terrorism problem faced by the Intelligence Community today. Based on her research on social networks, Sally is later hired by an agency in the Intelligence Community. However, Sally finds that things are a little more complicated in the Intelligence Community. First, there are multiple organizations and each maintains their own Intelligence Community databases (ICDB). Like the IMDB, these ICDBs provide an extensible set of relationships (is-member-of, suspected-role-in, also-known-as, communication-with, reported-by, first-identified-by) that describe and organize an universe of entities (individuals, terrorist cells, nations, political aims, terrorist events, financing sources, dates, etc.) However, each ICDB may impose a different organizational scheme on both the relationship types and the entities, and may use a different identifier for each relationship and entity. Also, the rules and procedures for intelligence collection also differ from organization to organization. As a consequence, the pragmatic meaning of the identified relationships and entities may differ in subtle collection-specific ways, leading to idiosyncratic issues when determining how an agent chooses to fuse these collections for a specific purpose. Other real world complications include: having multiple agents inside each organization that need to be able to update their ICDB; the pedigree of the information is critical; and agents in each organization need to be able to perform analysis on a fused view of the information base - as constrained by legal and procedural boundaries on what may be shared with whom and under what circumstances. The different ICDBs have evolved organically as internal applied R&D efforts. However, the Intelligence Community understands that these idiosyncratic solutions may be dated by recent developments in Semantic Web technologies and Sally is selected as part of a committee evaluating commercial "semantic web servers" for use as part of an ICDB solution. Ideally, a semantic web server would be able to speak and accept any of a number of semantic markup languages, and could be queried using any of a number of query languages. Further, Sally understands that an Intelligence Community semantic web server will contain 10E9 or more assertions - and a semantic web server must be able to process those requests efficiently. In an effort to make the information more useful while respecting the legal and procedural constraints on information sharing, Sally has been charged with exploring how semantic web servers can support privacy technologies, such as one-way hashes for information that can identify subjects, integrate with single-sign on architectures to facilitate constrained federated information views, maintain provenance information, and expose answers to simple yes/no questions without having to expose evidence and provenance on which those answers are based that could damage operators in the field or otherwise violate legal or procedural barriers on information sharing. -bryan
Received on Wednesday, 5 May 2004 07:51:58 UTC