Convolutional neural networks (CNNs) are the backbones of deep learning paradigms for numerous vision tasks. Early advancements in CNN architectures are primarily driven by human expertise and elaborate design. Recently, neural architecture search (NAS) was proposed with the aim of automating the network design process and generating task-dependent architectures. This paper introduces NSGA-Net -- an evolutionary search algorithm that explores a space of potential neural network architectures in three steps, namely, a population initialization step that is based on prior-knowledge from hand-crafted architectures, an exploration step comprising crossover and mutation of architectures, and finally an exploitation step that utilizes the hidden useful knowledge stored in the entire history of evaluated neural architectures in the form of a Bayesian Network. The integration of these components allows an efficient design of architectures that are competitive and in many cases outperform both manually and automatically designed architectures on CIFAR-10 classification task. The flexibility provided from simultaneously obtaining multiple architecture choices for different compute requirements further differentiates our approach from other methods in the literature.

No abstract

No abstract

No abstract

No abstract

No abstract

No abstract

Genetic algorithms have been used as an optimization tool using evolutionary strategies. Genetic algorithms cover three basic steps for population refinement selection, cross-over and mutation. In normal Real-coded genetic algorithm(RGA), the population of real variables generated after population refinement operations, is used for the computation of the objective function. In this paper we have shown the effect made by mapping the refined population towards better solutions and thereby creating more biased search. The mapping used is non-uniform in nature and is the function of the position of the individual w.r.t. the best solution obtained so far in the algorithm, and hence the name Non-Uniform RGA or in short NRGA. Tests were performed on standard benchmark problems. The results were promising and should encourage further research in this dimension.