Piscataway: Institute of Electrical and Electronics Engineers (IEEE)
摘要:
How to learn new knowledge without forgetting old knowledge is a key issue in designing an incremental-learning neural network. In this paper, we present a new incremental learning method for pattern recognition, called the “incremental backpropagation learning network”, which employs bounded weight modification and structural adaptation learning rules and applies initial knowledge to constrain the learning process. The viability of this approach is demonstrated for classification problems including the iris and the promoter domains
關聯:
IEEE Transactions on Neural Networks 7(3), pp.751-761