Structural Risk Minimization for Character Recognition

By Isabelle M Guyon, Vladimir Naumovich Vapnik, Bernhard E. Boser, L. Eon Bottou, Sara A. Solla. 1991

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, background): “, 1979), final prediction error (Akaike, 1970), generalized prediction error (Moody and Utans, 1994; Moody, 1992); see also (Holden, 1994; Wang et al., 1994; Amari and Murata, 1993; Wang et al., 1994; Guyon et al., 1992; Vapnik, 1992; Wolpert, 1994).”
On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, background): “2 M ’s Training To decrease bitsM + bitsH , we add a regularizing term to E, to punish excessive complexity [4, 5, 91, 294, 155, 135, 97, 170, 169, 104, 290, 7, 290, 87, 286, 319, 100, 102].”
The Effects of Adding Noise During Backpropagation Training on a Generalization Performance (1996, background): “Matsuoka proposes to minimize an objective function that is formed by adding a sensitivity measure to the standard error function. He then asserts that this objective function is minimized by the backpropagation algorithm with input noise. Using a recently proved convergence theorem on stochastic gr”
Probable networks and plausible predictions - a review of practical Bayesian methods for supervised neural networks (1995, methodology): “Recent work on `structural riskminimization' (Guyon et al. 1992) uses empirically tuned expressions of the form:Egen ' ED=N + c1 log(N= ) + c2N= (34)where is the `e ective V-C dimension' of the model, and is identical to the quantityin (23).”
A Tutorial on Support Vector Machines for Pattern Recognition (1998, background): “The problem which drove the initial development of SVMs occurs in several guises -the bias variance tradeoff (Geman, Bienenstock and Doursat, 1992), capacity control (Guyon et al., 1992), overfitting (Montgomery and Peck, 1992) -but the basic idea is the same.”
Local Learning Algorithms (1992, background): “The capacity control parameters y directly or indirectly define a subset W, of the weight space (Guyon et al. 1992).”
Thirty Years of Machine Learning: The Road to Pareto-Optimal Wireless Networks (2019, background): “Empirical risk minimization [11] and structural risk minimization [12] constitute a pair of fundamental strategies, where the latter can beneficially avoid the notorious “over-fitting” phenomenon.”
Flat Minima (1997, background): “(1994), Amari and Murata (1993), and Vapnik’s structural risk minimization (Guyon et al. 1992; Vapnik 1992).”
Gradient-Based Optimization of Hyperparameters (2000, background): “Increasing the penalty term (increasing the weight decay hyperparameter) corresponds to reducing the effective capacity ( Guyon, Vapnik, Boser, Bottou, & Solla, 1992 ) by forcing the solution to be in a zero-centered hypersphere of smaller radius, which may improve generalization.”
Bayesian Methods for Adaptive Models (2011, background): “Guyon et al.[30] have developed a measure of ‘effective V–C dimension’ of a regularised model.”