Combining the Predictions of Multiple Classifiers: Using Competitive Learning to Initialize Neural Networks

By Richard Maclin, Jude W. Shavlik. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, background): “…5.15, or hierarchical Kohonen nets (Section 5.6.4), to facilitate subsequent SL. Compare Stacked Generalization (Ting &Witten, 1997; Wolpert, 1992), and FNNs that profit from pre-training by competitive UL (e.g., Rumelhart & Zipser, 1986) prior to BP-based fine-tuning (Maclin & Shavlik, 1995).”
Diversity creation methods: a survey and categorisation (2004, background): “Maclin and Shavlik [ 40 ] present an approach to initializing neural network weights that uses competitive learning to create networks that are initialised far from the origin of weight space, thereby potentially increasing the set of reachable local minima; they show signican tly improved performan”
Extracting symbolic rules from trained neural network ensembles (2003, unknown): “Maclin and Shavlik [25] initialize individual networks at different points of the weight space.”
Ensembling neural networks: Many could be better than all (2002, background): “Opitz and Shavlik [34] exploit genetic algorithm to train diverse knowledge based component networks.”
Popular Ensemble Methods: An Empirical Study (1999, methodology): “…classifiers with dissimilar learning parameters (Alpaydin, 1993), different classifier architectures (Hashem, 1997), various initial neural-network weight settings (Maclin & Opitz, 1997; Maclin & Shavlik, 1995), or separate partitions of the training set (Breiman, 1996a; Krogh & Vedelsby, 1995).”
Lung cancer cell identification based on artificial neural network ensembles (2002, background): “Maclin and Shavlik [18] utilized competitive learning to generate individual networks and then combined their outputs via simple averaging.”
Model Extraction Warning in MLaaS Paradigm (2017, methodology): “In [10, 15], authors present techniques to combine NN models, which may be used by adversaries to collude for model extraction.”
Machine learning: a review of classification and combining techniques (2006, methodology): “For example, neural network techniques that have been employed include methods for training with different topologies, different initial weights and different parameters (Maclin and Shavlik 1995).”