Introduction to the theory of neural computation

By John A. Hertz, Anders Krogh, R. Palmer. 1994

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

On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models (2015, background): “Unlike feedforward NNs (FNNs; [95, 23]) and Support Vector Machines (SVMs; [287, 253]), RNNs can in principle interact with a dynamic partially observable environment in arbitrary, computable ways, creating and processing memories of sequences of input patterns [258].”
Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning (2016, methodology): “However, the complexity of computing the sensitivity is too high for large networks, so the magnitude-based pruning methods [6] use the magnitude of a weight to approximate its sensitivity; specifically, the small-magnitude weights are removed first.”
Structure-Adaptable Digital Neural Networks (1999, background): “This is achieved by means of a so called winner-take-all operation, which has been found to be biologically plausible [85].”
Information processing, dimensionality reduction and reinforcement learning in the basal ganglia. (2003, background): “However, this information is a scalar value, and is therefore evaluative, rather than instructive (Hertz et al., 1994).”
Comparative accuracies of artificial neural networks and discriminant analysis in predicting forest cover types from cartographic variables (1999, methodology): “Several recent textbooks describe the mechanics of ANNs (Hertz et al., 1991; Haykin, 1994; Masters, 1994; Bishop, 1995; Ripley, 1996).”
Neural Networks for Control (1993, background): “See the textbook [18] for a clear and well-written, if mathematically incomplete, introduction to neural nets.”
Artificial neural networks for computer-based molecular design. (1998, methodology): “The receptive ®elds of a fully trained UNN are comparable to the areas de®ned by Voronoi tessellation (or Dirichlet tessellation) of the input space (Hertz et al., 1991; Preparata and Shamos, 1985).”
Numerical implementation of a neural network based material model in finite element analysis (2004, background): “Details of NN training can be found in References [ 19, 20 ] .”