Learning internal representations by back-propagating errors

By D. Rumelhart. 1986

Subfields: Model Architectures, Optimization & Training

How researchers see this paper

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

In Model Architectures

In Optimization & Training

Citations in the corpus

Efficient Estimation of Word Representations in Vector Space (2013, methodology): “All models are trained using stochastic gradient descent and backpropagation [24].”
Natural Language Processing (Almost) from Scratch (2011, methodology): “Remark 6 (Modular Approach) The well known “back-propagation” algorithm (LeCun, 1985; Rumelhart et al., 1986) computes gradients using the chain rule.”
Deep, Narrow Sigmoid Belief Networks Are Universal Approximators (2008, methodology): “In discriminative applications the greedy algorithm is used to initialize the network’s parameters which are later fine-tuned using backpropagation [ 9 ] on a discriminative task.”
Reinforcement Learning with Factored States and Actions (2004, methodology): “The gradient was computed using error backpropagation (Rumelhart et al., 1986).”
Stochastic Configuration Networks: Fundamentals and Algorithms (2017, methodology): “The proposed learner model is generated incrementally by stochastic configuration (SC) algorithms, termed as Stochastic Configuration Networks (SCNs).”
Comparison of machine learning and traditional classifiers in glaucoma diagnosis (2002, background): “The MLP is the most popular architectures among other neural networks, such as the radial basis function [20], because it can be efficiently trained by error backpropagation [21].”
Generative models for discovering sparse distributed representations. (1997, methodology): “If an eecient algorithm can be found for tting models of this type it is likely to prove even more fruitful than the eecient backpropagation algorithm for multi-layer non-linear regression (Rumelhart et al., 1986).”
Extensions of recurrent neural network language model (2011, background): “As can be seen, 4-5 steps of BPTT training seems to be sufficient.”
Links between perceptrons, MLPs and SVMs (2004, methodology): “The rst one is the Multi-Layer Perceptron (MLP), which became useful with the introduction of the back-propagation training algorithm (LeCun, 1987; Rumelhart et al., 1986 ).”
Mutual information maximization: models of cortical self-organization. (1996, background): “For example, the back-propagation learning procedure [3] minimizes the mean squared error between the network’s output and the training signal.”