Curvature-driven smoothing: a learning algorithm for feedforward networks

By Charles M. Bishop. 1993

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): “Compare perturbation tolerance conditions (Bishop, 1993; Carter, Rudolph, & Nucci, 1990; Hanson, 1990; Kerlirzin & Vallet, 1993; Matsuoka, 1992; Minai & Williams, 1994; Murray & Edwards, 1993; Neti, Schneider, & Young, 1992).”
Bayesian Regularization and Pruning Using a Laplace Prior (1995, methodology): “For an approach to formal stabilization, more in the style of Tikhonov and Arsenin (1977), see Bishop (1993).”
Bayesian approach for neural networks--review and case studies (2001, methodology): “With an MLP model, minimizing the curvature (second derivative, Bishop, 1993) or training the derivatives to given target values (Lampinen & Selonen, 1997) leads to a rather complex treatise as the partial derivatives of the non-linear models depend on all the other inputs and weights.”
Neural networks for pattern recognition (1995, background): “i N /a n\(2) «^EEE(gf • ' n = l k % v * ' Derivatives of this regularizer with respect to the network weights can be found using an extended back-propagation algorithm (Bishop, 1993).”
Flat Minima (1997, unknown): “…8 ij : v ij = w ij g. Flatness condition 1 is a \robustness condition" (or \fault tolerance condition", or \perturbation tolerance condition" { see, e.g., Minai & Williams, 1994; Murray & Edwards, 1993; Neti et al., 1992; Matsuoka, 1992; Bishop, 1993; Kerlirzin & Vallet, 1993; Carter et al., 1990).”
Higher Order Contractive Auto-Encoder (2011, background): “Also in [4] the goal was to regularize a supervised objective, which is quite different from our focus on unsupervised feature learning.”
Preventing Over-Fitting during Model Selection via Bayesian Regularisation of the Hyper-Parameters (2007, background): “This regularisation term corresponds to curvature driven smoothing in multi-layer perceptron networks (Bishop, 1993), except that the model output ˆ y i is viewed as a function of the hyper-parameters, rather than of the weights.”
Pattern Recognition and Machine Learning (2006, unknown): “Derivatives of this regularizer with respect to the network weights can be found using an extended backpropagation algorithm (Bishop, 1993).”