DG Researchers Parsing in Tongues
José Fortes, Professor and BellSouth Eminent Scholar at the University of Florida, is principal investigator on one of the rare international Digital Government projects. He has dealt with many of the same issues - the difficulties of integrating heterogeneous databases and different agency rules - as his DG counterparts on US-only projects.
But he has also encountered a unique set of challenges, not least of which is that not all participants on this project share a common language.
Fortes and co-PI Stanley Su, Director of the Database Systems Research and Development Center, University of Florida are part of a multi-national team of collaborators that includes faculty from: Carnegie Mellon, the University of Colorado, North Carolina State University, the University of Massachusetts and the University of Belize and the Pontificia Universidad Catolica Madre y Maestra of the Dominican Republic.
The academics are working with officials in the Belize and Dominican Republic governments, as well as the Office of Science and Technology and the Inter-American Drug Abuse Control Commission of the Organization of American States, on the problem of cross-border drug control. Their joint task is to contribute to the automation of the border control process to improve its effectiveness and facilitate statistics for CICAD's Multilateral Evaluation Mechanism (MEM).
Entering and sharing data are the first steps towards tracking the movements of drug-dealers across international boarders, says Fortes. "The MEM requires that countries collect, share and analyze extensive amounts of information in a standard format," he says. The extra difficulty lies in integrating data from different countries in two languages - English and Spanish.
"There are issues of a socio-political nature that have nothing to do with technology," says Fortes. "Individual countries differ in their languages, laws, regulations, cultures, administrative structures, resources, geopolitical characteristics and stages of technological development."
The last of these aspects provides one of the toughest challenges.
Law enforcement officials in remote areas with insufficient funding may have access to nothing more technologically sophisticated than a telephone. However beautifully designed, any interface based on a computer screen would be completely inaccessible. The team needed to design a system that could integrate data entered in different formats using a conversational style and have it translated automatically where necessary. Such a system could, in the future, support the automatic conversion of voice into text, so that the results could be received by telephone.
It would seem like it would take an entire computer science department to pull off such a challenge and it very nearly did. The university researchers include experts on speech-based interfaces, machine translation, databases, information retrieval, Internet-computing, software requirements and networking. The government partners include experts in various areas of law enforcement and international diplomacy.
Obviously, it's an extraordinarily complex collaboration, so much so that Fortes is only half-joking when he says, "The social-cultural interactions could become a case study for another Digital Government researcher. Belize has a British background, the Dominican Republic has a Spanish background and this all comes across when working with different agencies."
Ruth Marie Connolly, Principal Specialist in the OAS' Office of Science and Technology, says, "The Organization of American States (OAS) has received from its 35 Western Hemisphere states various mandates to develop digital government and telecommunications in the Hemisphere. Therefore this pilot project, which paves the way for more extensive collaboration in this area is a very important initiative and a test bed for future and more extensive national and regional efforts."
"There are also differences in time - not time-zones, as you might expect, but timing within agencies and the timing of national elections," says Fortes. "Your government collaborators could be gone, having left office under a new administration - or they could simply be on a holiday the other countries don't observe."
Yet despite the odds and obstacles, the team has produced a prototype system that is being deployed in tests at the collaborating universities in the US and Belize. (The US Embassy in Belize has donated computers and other equipment for the project.) Demonstrations and deployment in the Dominican Republic are scheduled for the second half of 2004.
Border control agencies also will beta-test the system, providing feedback to the research team for further development. The researchers say they hope to have a fully operational prototype in place by 2005. This prototype does not yet support voice-to-text conversions, only textual conversational interfaces. Funding is being sought for this and other extensions that might allow the OAS to disseminate the system to other countries.
Fortes offers this advice for anyone contemplating an equally ambitious idea:
"Differences that have nothing to do with what you want to put in the project should be minimized - they just get in the way. There should be a preference for small and well-defined projects within the larger, overall scope. One very useful mechanism for doing this should be to start with an exploratory grant to do the groundwork needed to engage international partners and clearly define the focus of the project."
|This site is maintained by the Digital Government Research Center at the University of Southern California's Information Sciences Institute.||" CONTACT " POLICIES|