Field Theoretical Analysis of On-Line Learning of Probability Distributions
- 25 October 1999
- journal article
- research article
- Published by American Physical Society (APS) in Physical Review Letters
- Vol. 83 (17) , 3554-3557
- https://doi.org/10.1103/physrevlett.83.3554
Abstract
On-line learning of probability distributions is analyzed from the field theoretical point of view. We can obtain an optimal on-line learning algorithm, since a renormalization group enables us to control the number of degrees of freedom of a system according to the number of examples. We do not learn parameters of a model, but probability distributions themselves. Therefore, the algorithm requires no a priori knowledge of a model.Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Field Theories for Learning Probability DistributionsPhysical Review Letters, 1996
- On-line versus Off-line Learning from Random Examples: General ResultsPhysical Review Letters, 1996
- On-line Gibbs LearningPhysical Review Letters, 1996
- Bounds for Predictive Errors in the Statistical Mechanics of Supervised LearningPhysical Review Letters, 1995
- Learning from noisy data: An exactly solvable modelPhysical Review E, 1995
- Local and Global Convergence of On-Line LearningPhysical Review Letters, 1995
- Network information criterion-determining the number of hidden units for an artificial neural network modelIEEE Transactions on Neural Networks, 1994
- Statistical Theory of Learning Curves under Entropic Loss CriterionNeural Computation, 1993
- Statistical mechanics of learning from examplesPhysical Review A, 1992
- Differential-Geometrical Methods in StatisticsPublished by Springer Nature ,1985