Search

Training artificial neural networks requires a tedious empirical evaluation to determine a suitable neural network architecture. To avoid this empirical process several techniques have been proposed to automatise the architecture selection process. In this paper, we propose a method to perform parameter and architecture selection for a quantum weightless neural network (qWNN). The architecture selection is performed through the learning procedure of a qWNN with a learning algorithm that uses the principle of quantum superposition and a non-linear quantum operator. The main advantage of the proposed method is that it performs a global search in the space of qWNN architecture and parameters rather than a local search. [ABSTRACT FROM AUTHOR]

In order to investigate the dynamics of a quantum weightless neuron node we feed its output back as input. Due to the fact that controlled operators used in the neuron circuit usually generate entanglement, we propose a mathematical method to extract the output at time t and build from that output the input at time t + 1 . As a result the time evolution is a real-valued nonlinear map with one real parameter. The dynamics orbits are plotted showing acute sensitivity to initial conditions clearly exhibiting nonlinearity by just looking at amplitude graphs. The fractal geometry and regions of convergence are discussed by their Julia Set images and a new measure for model comparison is put forward. [ABSTRACT FROM AUTHOR]