Scientific Problem Types and Computational Discovery Systems
Scientific research involves an intricate set of interrelated problems. The problems range from data and observation to technique and experimentation to theory formation and revision. A schematic analysis of these problem types, developed by Root-Bernstein ( 1989), is presented in Table 10.1.
To what extent have artificial intelligence approaches ( Boden, 1996) to the problems of scientific discovery encountered and solved the types of problems described by Root-Bernstein ( 1989)? I have constructed Table 10.2, "Scientific Problem Types and Computational Discovery Systems," as a parallel to Root-Bernstein's table. Table 10.2 suggests that computational discovery systems have made considerable progress in dealing with the types of problems encountered in standard scientific research.