Researchers who make theoretical advances also need some way to demonstrate that an advance really does have general, overall positive consequences for system performance. For this it is necessary to evaluate the system on a set of problems that is sufficiently large and diverse to be somehow representative of the intended application area as a whole. It is only a small step from system evaluation to a communal system competition. The CADE ATP System Competition (CASC) has been run annually since 1996. Any competition is difficult to design and organize in the first instance, and to then run over the years. In order to obtain the full benefits of a competition, a thoroughly organized event, with an unambiguous and motivated design, is necessary. For some issues relevant to the CASC design, inevitable constraints have emerged. For other issues there have been several choices, and decisions have had to be made. This paper describes the evolution of CASC, paying particular attention to its design, design changes, and organization.
|Original language||English (US)|
|Number of pages||12|
|State||Published - Aug 14 2002|
- Automated theorem proving
ASJC Scopus subject areas
- Artificial Intelligence