A Constrained Optimization Approach to Preserving Prior Knowledge During Incremental Training

In this paper, a supervised neural network training technique based on constrained optimization is developed for preserving prior knowledge of an input

• Type: Journal
• Year: 2008
• Where: TNN
• Authors: Silvia Ferrari, Mark Jensenius
• Keywords: Equality Constraints, Incremental Training, Prior Knowledge, TNN 2008