Computer Aided Design of Learning Systems in PROLOG

摘要

The paper describes an approach to computer aided design of symbol based machine learning programs in Prolog. By embedding Prolog clauses of previously prepared programs with special instructions a generator is able to generate fully functional solutions according to the data passed to it. The paper shows why such generation might be useful for machine learning programs. The approach is compared with other works in the field of automatic generation of logic programs. The use of some symbolic learning methods in different areas of AI is also described.