Applications

CIRL is fundamentally a theoretical research laboratory; our goal is to make scientific contributions to the AI enterprise. We recognize, however, that a theoretical development that is attractive on paper may be all but useless in practice. We are therefore committed to not just arguing that our ideas are of value, but to showing that they are. This commitment to demonstration causes work at CIRL to proceed in three phases:

First, an interesting theoretical idea is developed. The idea may extend existing representational mechanisms with little or no computational cost, or may improve the performance of our existing algorithms.
Next, the idea is incorporated into a general-purpose tool. This is something of a "sanity check," and serves to demonstrate that the attractive theoretical properties of the idea can be realized in practice.
The general-purpose tools currently available include general reasoning tools ("satisfiability"), along with general tools for problem solving in the areas of planning and workflow.
Finally, the implemented idea is applied to a realistic problem of practical interest. The application may be based on a general-purpose tool, or on a specialized tool in one of our existing areas of expertise or a new such area. Our existing areas of expertise include bridge, scheduling, and route planning.

Subareas:

General-purpose tools

Problem domains

Back to the CIRL Overview Page.