Advanced supervised learning in multi-layer perceptrons - From backpropagation to adaptive learning algorithms

Computer Standards and Interfaces

Volumn 16, Issue 3, 1994, Pages 265-278

  Riedmiller, Martin ^a  

^a UNIVERSITY OF KARLSRUHE   (Germany)

Author keywords

Adaptive learning algorithms; Benchmark problems; Feedforward networks; Multi layer perceptrons; Robustness; Supervised learning

Indexed keywords

ADAPTIVE SYSTEMS; ALGORITHMS; CLASSIFICATION (OF INFORMATION); CONVERGENCE OF NUMERICAL METHODS; LEARNING SYSTEMS; OPTIMIZATION; STANDARDS;

ADAPTIVE LEARNING ALGORITHMS; BACKPROPAGATION LEARNING; BENCHMARK PROBLEMS; FEEDFORWARD NETWORKS; GRADIENT DESCENT TECHNIQUE; MULTILAYER PERCEPTRONS; SCALING PROPERTIES; SUPERVISED LEARNING;

NEURAL NETWORKS;

EID: 0028466750     PISSN: 09205489     EISSN: None     Source Type: Journal    
DOI: 10.1016/0920-5489(94)90017-5     Document Type: Article

References (12)

1. Learning internal representations by error propagation
2. Improved convergence rate of backpropagation with dynamic adaptation of the learning rate
3. A scaled conjugate gradient algorithm for fast supervised learning
4. Optimization of the backpropagation algorithm for training multilayer perceptrons
5. Efficient parallel learning algorithms for neural networks
6. Increased rates of convergence through learning rate adaptation
7. Supersab Fast adaptive backpropagation with good scaling properties
8. Speeding up backpropagation
9. An empirical study of learning speed in back-propagation networks
10. A direct adaptive method for faster backpropagation learning: The RPROP algorithm
11. Learning to tell two spirals apart