On complexity analysis of supervised MLP-learning for algorithmic comparisons

@inproceedings{Mizutani2001, abstract = {This paper presents the complexity analysis of a standard supervised MLP-learning algorithm in conjunction with the well-known backpropagation, an efficient method for evaluation of derivatives, in either batch or incremental learning mode. In particular, we detail the cost per epoch (i.e., operations required for processing one sweep of all the training data) using \“approximate\” FLOPs (floating point operations) in a typical backpropagation for solving neural networks nonlinear least squares problems. Furthermore, we identify erroneous complexity analyses found in the past NN literature. Our operation-count formula would be very useful for a given MLP architecture to compare learning algorithms}, author = {Mizutani, E. and Dreyfus, S. E. }, booktitle = {Neural Networks, 2001. Proceedings. IJCNN '01. International Joint Conference on}, citeulike-article-id = {2919569}, doi = {10.1109/IJCNN.2001.939044}, journal = {Neural Networks, 2001. Proceedings. IJCNN '01. International Joint Conference on}, pages = {347--352 vol.1}, posted-at = {2008-06-23 19:42:46}, priority = {2}, title = {On complexity analysis of supervised MLP-learning for algorithmic comparisons}, url = {http://dx.doi.org/10.1109/IJCNN.2001.939044}, volume = {1}, year = {2001} }

TY - CONF ID - Mizutani2001 L3 - citeulike-article-id:2919569 TI - On complexity analysis of supervised MLP-learning for algorithmic comparisons T2 - Neural Networks, 2001. Proceedings. IJCNN '01. International Joint Conference on JF - Neural Networks, 2001. Proceedings. IJCNN '01. International Joint Conference on VL - 1 SP - 347 EP - 352 vol.1 N2 - This paper presents the complexity analysis of a standard supervised MLP-learning algorithm in conjunction with the well-known backpropagation, an efficient method for evaluation of derivatives, in either batch or incremental learning mode. In particular, we detail the cost per epoch (i.e., operations required for processing one sweep of all the training data) using “approximate” FLOPs (floating point operations) in a typical backpropagation for solving neural networks nonlinear least squares problems. Furthermore, we identify erroneous complexity analyses found in the past NN literature. Our operation-count formula would be very useful for a given MLP architecture to compare learning algorithms AU - Mizutani, E AU - Dreyfus, SE PY - 2001/// UR - http://dx.doi.org/10.1109/IJCNN.2001.939044 ER -