Keynote address - data abstraction and hierarchy
OOPSLA '87 Addendum to the proceedings on Object-oriented programming systems, languages and applications (Addendum)
What size net gives valid generalization?
Neural Computation
Generalization by weight-elimination with application to forecasting
NIPS-3 Proceedings of the 1990 conference on Advances in neural information processing systems 3
Parallel structured networks for solving a wide variety of matrix algebra problems
Journal of Parallel and Distributed Computing - Special issue on neural computing on massively parallel processing
Neural network approach to computing matrix inversion
Applied Mathematics and Computation
Neural networks and the bias/variance dilemma
Neural Computation
Neural Computation
A practical Bayesian framework for backpropagation networks
Neural Computation
A recurrent neural network for real-time matrix inversion
Applied Mathematics and Computation
Some new results on neural network approximation
Neural Networks
Experiments on the transfer of knowledge between neural networks
Proceedings of a workshop on Computational learning theory and natural learning systems (vol. 1) : constraints and prospects: constraints and prospects
Artificial intelligence: a modern approach
Artificial intelligence: a modern approach
Adaptive critic for sigma-pi networks
Neural Networks
Assessing generalization of feedforward neural networks
Assessing generalization of feedforward neural networks
Knowledge reuse in multiple classifier systems
Pattern Recognition Letters - special issue on pattern recognition in practice V
Neural Networks: A Comprehensive Foundation
Neural Networks: A Comprehensive Foundation
An Introduction to Neural Networks
An Introduction to Neural Networks
Neural Networks for Pattern Recognition
Neural Networks for Pattern Recognition
The Handbook of Mathematics and Computational Science
The Handbook of Mathematics and Computational Science
Artificial Intelligence: A Guide to Intelligent Systems
Artificial Intelligence: A Guide to Intelligent Systems
Symmetry in class and type hierarchy
CRPIT '02 Proceedings of the Fortieth International Conference on Tools Pacific: Objects for internet, mobile and embedded applications
Learning Probabilistic RAM Nets Using VLSI Structures
IEEE Transactions on Computers
Symmetry Breaking in Software Patterns
GCSE '00 Proceedings of the Second International Symposium on Generative and Component-Based Software Engineering-Revised Papers
An Overview of Hybrid Neural Systems
Hybrid Neural Systems, revised papers from a workshop
Improving the Generalization Capability of the Binary CMAC
IJCNN '00 Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks (IJCNN'00)-Volume 3 - Volume 3
Learning in networks of structured hypercubes
Learning in networks of structured hypercubes
How initial conditions affect generalization performance in large networks
IEEE Transactions on Neural Networks
Generalization properties of modular networks: implementing the parity function
IEEE Transactions on Neural Networks
IEEE Transactions on Neural Networks
Pruning recurrent neural networks for improved generalization performance
IEEE Transactions on Neural Networks
| Hi-index | 0.01 |
Generalization, in its most basic form, is an artificial neural network's (ANN's) ability to automatically classify data that were not seen during training. This paper presents a framework in which generalization in ANNs is quantified and different types of generalization are viewed as orders. The ordering of generalization is a means of categorizing different behaviours. These orders enable generalization to be evaluated in a detailed and systematic way. The approach used is based on existing definitions which are augmented in this paper. The generalization framework is a hierarchy of categories which directly aligns an ANN's ability to perform table look-up, interpolation, extrapolation, and hyper-extrapolation tasks. The framework is empirically validated. Validation is undertaken with three different types of regression task: (1) a one-to-one (o-o) task, f(x):x"i-y"j; (2) the second, in its f(x):{x"i,x"i"+"1,...}-y"j formulation, maps a many-to-one (m-o) task; and (3) the third f(x):x"i-{y"j,y"j"+"1,...} a one-to-many (o-m) task. The first and second are assigned to feedforward nets, while the third, due to its complexity, is assigned to a recurrent neural net. Throughout the empirical work, higher-order generalization is validated with reference to the ability of a net to perform symmetrically related or isomorphic functions generated using symmetric transformations (STs) of a net's weights. The transformed weights of a base net (BN) are inherited by a derived net (DN). The inheritance is viewed as the reuse of information. The overall framework is also considered in the light of alignment to neural models; for example, which order (or level) of generalization can be performed by which specific type of neuron model. The complete framework may not be applicable to all neural models; in fact, some orders may be special cases which apply only to specific neuron models. This is, indeed, shown to be the case. Lower-order generalization is viewed as a general case and is applicable to all neuron models, whereas higher-order generalization is a particular or special case. This paper focuses on initial results; some of the aims have been demonstrated and amplified through the experimental work.