Efficient Learning Through Evolution: Neural Programming and Internal Reinforcement

Abstract

Genetic programming (GP) can learn complex concepts by searching for the target concept through evolution of population of candidate hypothesis programs. However, unlike some learning techniques, such as Artificial neural networks (ANNs), GP does not have a principled procedure for changing parts of a learned structure based on that structure's performance on the training data. GP is missing a clear, locally optimal update procedure, an equivalent of gradient-descent backpropagation for ANNs. This article introduces a new mechanism, "internal reinforcement, " for defining and using performance feedback on program evolution. A new connectionist representation for evolving parameterized programs, "neural programming" is also introduced. We present the algorithms for the generation of credit and blame assignment in the process of learning programs using neural programming and internal reinforcement. The article includes some of our extensive experiments that demonstr...