The Data Center

AN OVERVIEW OF SEMANTIC MODELING

In every organization, there are hundreds of applications for mathematical models. However, models are seldom shared or re-used because of a lack of interconnectivity between separate models and between models and data. This lack of interconnectivity increases development time for a model in practice and limits overall productivity.

To overcome this practical problem requires a new approach that involves creating an interoperable modeling environment through the use of computer languages and protocols. The next step for computing is some sort of network wide insight or awareness that presently does not exist with functions like email, documentation, and data storage. In a way similar to linking physical objects to the Internet (Auto-ID), The Data Center is developing methods to link abstractions (different types of models) together using state of the art computational techniques.

This will become even more important now that common standards have been established for Auto-ID technology. As sensors become a practical tool in supply chains, commercial, military, and healthcare applications there will be a flood of new data to be analyzed. The Data Center provides the means to analyze this data, leading to new insights. Taking full advantage of the increased availability of data is only possible through networks, standards, languages, and protocols that organize the modeling environment.

The MIT Data Center participates the development of new computer languages and protocols, incorporated into a software package that will allow for the linking of models and for distributed computing. This will provide the ability for developers to post models for others to download and use, potentially for a fee. A special "semantic" search engine will facilitate this process to allow:

a. The sharing and recombination of model elements through an intelligent network

b. The operation of models in parallel on separate computing platforms

c. A repository of modes that can be applied across disciplines

Formal research in Semantic Modeling began on January 1, 2005. Further information about The MIT Data Project can be obtained from Dave Brock, dlb@mit.edu or Ed Schuster at edmund_w@mit.edu.