The Upstart Algorithm: A Method for Constructing and Training Feedforward Neural Networks
Marcus R. Frean · Neural Computation · 1990
A general method for building and training multilayer perceptrons composed of linear threshold units is proposed. A simple recursive rule is used to build the structure of the network by adding units as they are needed, while a modified perceptron algorithm is used to learn the connection strengths. Convergence to zero errors is guaranteed for any boolean classification on patterns of binary variables. Simulations suggest that this method is efficient in terms of the numbers of units constructed, and the networks it builds can generalize over patterns not in the training set.
Cited by 512 works. doi:10.1162/neco.1990.2.2.198
Authors
References
- Learning Internal Representations by Error Propagation
- Principles of Neurodynamics.
- Large Automatic Learning, Rule Extraction, and Generalization.
- Monte Carlo Methods in Statistical Physics
- Learning in feedforward layered networks: the tiling algorithm
- Tilinglike learning in the parity machine
- Learning the Unlearnable