Conceptual science and beyond physical science Pdf

Conceptual science is an emerging discipline which merges the fields of cognitive neuroscience, philosophy, psychology, artificial intelligence, anthropology and computer science to study the nature of concepts and representations. To date there is no a unified theory in this area. This book presents methods and explains main concepts based on Artificial Neural Networks (ANNs). The main objective is to explore the logical systems. Inspired by those studies we have performed research about learning machines using ANS whose structure was generated using genetic programming (GP). The sample conceptual networks were generated by subjecting semantic trees evolved by GP to operations of genetic pruning and reshaping operations. Additional analytical tools were used to study how the logical structure influences learning and generalization. In this framework our network presents interesting properties. We observe that every part of the network is able to learn subconcepts by itself, so not a single neuron will be left without learning from data being fed into it. At the same time a new concept is obtained from data outliers. Also for such concepts separate neurons are capable of learning and generalizing. This is a result of usage ANS in GP technique as part of improving learning efficiency increasing their ability to represent novel concepts at a faster pace towards convergence, hence prevents overfitting and

Conceptual science is the study of scientific ideas and theories and how they relate to each other and modern beliefs. It tries to answer questions regarding science, technology and mathematics that do not have empirical evidence. For example, we can still study the information of physics without having to experiment with physical objects or create hypotheses that are based on real life.

An introduction to conceptual science, a freely available course by author and scientist James S. Trefil Ph.D. This course includes the following lectures: