Adaptive resonance theory (ART) algorithms represent a class of neural network architectures which self-organize stable recognition categories in response to arbitrary sequences of input patterns. The authors discuss incorporation of supervision into one of these architectures, ART2. Results of numerical experiments indicate that this new semi-supervised version of ART2 (SMART2) outperformed ART for classification problems. The results and analysis of runs on several data sets by SMART2, ART2, and backpropagation are analyzed. The test accuracy of SMART2 was similar to that of backpropagation. However, SMART2 network structures are easier to interpret than the corresponding structures produced by backpropagation.
C. J. Merz et al., "SeMi-Supervised Adaptive Resonance Theory (SMART2)," Proceedings of the International Joint Conference on Neural Networks, 1992, Institute of Electrical and Electronics Engineers (IEEE), Jan 1992.
The definitive version is available at https://doi.org/10.1109/IJCNN.1992.227046
International Joint Conference on Neural Networks, 1992
Keywords and Phrases
ART2; SMART2; Adaptive Resonance Theory; Classification Problems; Neural Nets; Neural Network Architectures; Pattern Recognition; Self Organizing; Stable Recognition Categories; Unsupervised Learning
Article - Conference proceedings
© 1992 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.