Evolving ARTMAP neural networks using Multi-Objective Particle Swarm Optimization

In this paper, a supervised learning strategy based on a Multi-Objective Particle Swarm Optimization (MOPSO) is introduced for ARTMAP neural networks. It is based on the concept of neural network evolution in that particles of a MOPSO swarm (i.e., network solutions) seek to determine user-defined parameters and network (weights and architecture) such that generalisation error and network resources are minimized. The performance of this strategy has been assessed with fuzzy ARTMAP using synthetic and real-world data for video-based face classification. Simulation results indicate that when the MOPSO strategy is used to train fuzzy ARTMAP, it produces a significantly lower classification error than when trained using standard hyper-parameter settings. Furthermore, the non-dominated MOPSO solutions represent a better compromise between error and resource allocation than mono-objective PSO-based strategies that minimizes only classification error. Overall, results obtained with the MOPSO strategy reveal the importance of optimizing parameters and network for each problem, where both error and resources are minimized during fitness evaluation.

I. Introduction

In batch supervised learning mode, fuzzy ARTMAP [1] may be efficient in that its asymptotical generalisation error can be achieved for a moderate time and memory complexity. As such, they have been successfully applied to a wide variety of pattern recognition problems such as the recognition of radar signals and handwritten characters, multi-sensor image fusion, remote sensing and data mining, and biometric authentification (with signature, face, etc.) [22]. Nonetheless, a drawback of fuzzy ARTMAP is its inability to learn decision boundaries between class distributions that consistently yield low generalization error [11].