Summary: INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS1
Int. J. Intel. Systems XX, yyy­zzz (1999)
© 1999 John Wiley & Sons, Ltd Int. J. Intel. Systems. XX, yyy­zzz (1999)
CONSTRUCTIVE REINFORCEMENT LEARNING
JOSE HERNANDEZ­ORALLO *
Department of Information Systems and Computation, Technical University of Valencia, Camí de Vera 14, Aptat. 22.012
E­46071, Valencia, Spain
ABSTRACT
This paper presents an operative measure of reinforcement for constructive learning methods, i.e., eager
learning methods using highly expressible (or universal) representation languages. These evaluation tools al­
low a further insight in the study of the growth of knowledge, theory revision and abduction. The final ap­
proach is based on an apportionment of credit wrt. the `course' that the evidence makes through the learnt the­
ory. Our measure of reinforcement is shown to be justified by cross­validation and by the connection with
other successful evaluation criteria, like the MDL principle. Finally, the relation with the classical view of re­
inforcement is studied, where the actions of an intelligent system can be rewarded or penalised, and we discuss
whether this should affect our distribution of reinforcement. The most important result of this paper is that the
way we distribute reinforcement into knowledge results in a rated ontology, instead of a single prior distribu­
tion. Therefore, this detailed information can be exploited for guiding the space search of inductive learning
algorithms. Likewise, knowledge revision may be done to the part of the theory which is not justified by the
evidence. © XXXX John Wiley & Sons, Ltd.

Source: Alpuente, María - Departamento de Sistemas Informáticos y Computación, Universitat Politècnica de València
Hernández Orallo, José - Departamento de Sistemas Informáticos y Computación, Universitat Politècnica de València

Collections: Computer Technologies and Information Sciences