Searching for optimal MLP

Backpropagation based on minimization algorithms is replaced by heuristic search techniques for quantized weights. The resulting algorithm is fast, avoids local minima of the cost function, and may be used either as initialization method for standard backpropagation or as a logical rule extraction technique.

[1]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[2]  Jerzy Korczak,et al.  Optimization and global minimization methods suitable for neural networks , 1999 .

[3]  Nada Lavrac,et al.  The Multi-Purpose Incremental Learning System AQ15 and Its Testing Application to Three Medical Domains , 1986, AAAI.

[4]  Włodzisław Duch,et al.  Methodology of extraction , optimization and application of logical rules , 1999 .

[5]  Wlodzislaw Duch,et al.  Statistical Methods for Construction of Neural Networks , 1998, ICONIP.

[6]  Norbert Jankowski,et al.  Initialization and optimization of multilayered perceptrons , 1997 .

[7]  Vipin Kumar,et al.  Search in Artificial Intelligence , 1988, Symbolic Computation.

[8]  Wlodzislaw Duch,et al.  Extraction of crisp logical rules using constrained backpropagation networks , 1997, ESANN.

[9]  Sholom M. Weiss,et al.  An Empirical Comparison of Pattern Recognition, Neural Nets, and Machine Learning Classification Methods , 1989, IJCAI.

[10]  Ivan Bratko,et al.  ASSISTANT 86: A Knowledge-Elicitation Tool for Sophisticated Users , 1987, EWSL.

[11]  Włodzisław Duch.  ALTERNATIVES TO GRADIENT-BASED NEURAL TRAINING , 1999 .

[12]  Larry J. Eshelman,et al.  Using Weighted Networks to Represent Classification Knowledge in Noisy Domains , 1988, ML.

[13]  Krzysztof Grabczewski,et al.  Extraction of logical rules from backpropagation networks , 1998 .

[14]  Wlodzislaw Duch,et al.  Logical Rules for Classification of Medical Data Using Ontogenic Neural Algorithm , 1997, ICONIP.