(Enter summary)
Abstract: Ideally pattern recognition machines provide constant output when
the inputs are transformed under a group G of desired invariances.
These invariances can be achieved by enhancing the training data
to include examples of inputs transformed by elements of G, while
leaving the corresponding targets unchanged. Alternatively the
cost function for training can include a regularization term that
penalizes changes in the output when the input is transformed under
the group.
This paper relates... (Update)
Context of citations to this paper: More
.... 6] 7] Their use has been justified through regularization theory in connection with their energy minimization properties [8] 9] [10]. A radial basis function approximation in p dimensions (x # R p ) has the generic form f(x) X k#Z a k #(#x x k #) 1) where...
...adding the virtual examples L t x i in the training set. Indeed the two approaches are related and some equivalence can be shown [6]. So why not just add virtual examples This is the idea of the Virtual Support Vector (VSV) method [10] The reason is the following: if a...
Cited by: More
Anisotropic Noise Injection for Input Variables Relevance.. - Grandvalet (2000)
(Correct)
On a Kernel-Based Method for Pattern Recognition.. - Smola, Schölkopf (1998)
(Correct)
From Samples to Objects in Kernel Methods - Pozdnoukhov (2003)
(Correct)
Active bibliography (related documents): More All
0.8: From Data Distributions to Regularization in Invariant Learning - Leen (1995)
(Correct)
0.1: Adaptive Averaging in Higher Order Neural Networks for Invariant.. - Kröner (1995)
(Correct)
0.1: Learning Algorithm for Structured Invariant Neural Networks - Kröner (1996)
(Correct)
Similar documents based on text: More All
0.1: Networks, volume IV, pages 589 -- 594. [Yu et al., 1990].. - Mexico Volume Speech
(Correct)
0.1: From Mixtures of Mixtures to Adaptive Transform Coding - Cynthia Archer And (2001)
(Correct)
0.1: Bibliography - Zhao An
(Correct)
Related documents from co-citation: More All
8: The Nature of Statistical Learning Theory (context) - Vapnik - 1995
8: Self-Organizing Maps (context) - Kohonen - 1995
8: Training with noise is equivalent to Tikhonov regularization
- Bishop - 1995
BibTeX entry: (Update)
Todd K. Leen. From data distributions to regularization in invariant learning. Neural Computation, 3(1):135--143, 1991. http://citeseer.ist.psu.edu/leen95from.html More
@inproceedings{ leen95from,
author = "Todd K. Leen",
title = "From Data Distributions to Regularization in Invariant Learning",
booktitle = "Advances in Neural Information Processing Systems",
volume = "7",
publisher = "The {MIT} Press",
editor = "G. Tesauro and D. Touretzky and T. Leen",
pages = "223--230",
year = "1995",
url = "citeseer.ist.psu.edu/leen95from.html" }
Citations (may not include all citations):
57
Training with noise is equivalent to Tikhonov regularization
- Bishop - 1995
45
Handwritten digit recognition with a back-propagation networ..
- Le Cun, Boser et al. - 1990
30
Lie Groups and Algebras with Applications to Physics (context) - Sattinger, Weaver - 1986
13
A method for learning from hints (context) - Abu-Mostafa - 1993
6
Encoding geometric invariances in higherorder neural network.. (context) - Giles, Griffin et al. - 1988
2
Tanget prop - a formalism for specifying selected invariance.. (context) - Simard, Victorri et al. - 1992
The graph only includes citing articles where the year of publication is known.
Documents on the same site (http://cse.ogi.edu/~tleen/publications.html): More
Exact and Perturbation Solutions for the Ensemble Dynamics - Leen (1998)
(Correct)
From Data Distributions to Regularization in Invariant Learning - Leen (1995)
(Correct)
Multi-Stream Video Fusion Using Local Principal Components.. - Sharma, Pavel, Leen
(Correct)
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC