RealTime IT News

DARPA Pads Semantic Web Contract

Web software developer Teknowledge Tuesday said it has won an extended government contract to help build an evolving version of the World Wide Web that centers on the meaning of words.

The Palo Alto, Calif.-based firm said the Defense Advanced Research Projects Agency (DARPA) has added $634,057 to its now $1.7 million budget to help build the DARPA Agent Markup Language (DAML).

The contract is centered on developing an Agent Semantic Communication Service, which lets users access relevant data through an XML framework. DARPA calls it the "Semantic Web".

For example, when you tell a person something, he can combine the new fact with an old one and tell you something new. When you tell a computer something in XML, it may be able to tell you something new in response, but only because of some other software it has that's not part of the XML spec. DARPA is using the new language to create a program that assigns similar semantics to a "subProperty" tag.

DARPA, which was responsible for funding much of the development of the Internet we know today, has been working on the DAML spec since August 2000 as a way to augment the Web and improve data mining. Currently, the agency is working with the W3C through a various working groups to implement it.

Teknowledge says the open source framework will be one of the tools that drives the Semantic Web from research vision to practical reality.

"[Our] Knowledge Systems group is doing extremely innovative research and development that will provide a semantic foundation for the next generation of financial and security software applications," Teknowledge president and CEO Neil Jacobstein said in a statement. "The software in our lives today may be graphical, fast, and useful, but it knows very little about its users tasks, or the underlying meaning of what is displayed on a computer screen. This project is exciting because it offers the possibility of building software that can begin to provide meaningful support for user tasks. I would not describe this as a breakthrough, but rather an incremental step in a long development path. This kind of software has been on Teknowledge's radar screen for over twenty years, and its day is coming."

The technology is steeped in Teknoledge's Suggested Upper Merged Ontology (SUMO) and the company says it has mapped it to over 100,000 word senses in the WordNet natural language lexicon. The company says its research helps give technical users a comprehensive language for asking precise questions.

"Our software allows users to get answers to questions that have been precisely defined, rather than just a sampling of documents that contain relevant keywords, as in most current Web searches," said Teknowledge Director of Knowledge Systems Adam Pease. "We can also perform inference during search that allows users to get answers that are not literally on the Web, but must be inferred by the computer software."

Pease said the next step in this process is to provide a simplified language to support non-technical end users.

Spawned in the early 1980s at the Stanford University computer science labs, the Teknowledge's software is focused improves business and transaction processes for government agencies and financial services companies. Banks use its TekPortal software (part of its Financial Services unit) to aggregate and mine customer data, and to provide money and financial management tools to investors. Teknowledge also provides software development and services for network security, Web-based training, and business intelligence through its Security, Training, Distributed Systems, and Knowledge Systems divisions. Clients include the US Air Force, the US Navy, Fiserv, and NetBank.