Modeles connexionnistes de l'apprentissage

By Yann LeCun, F. Soulié. 1987

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

A practical tutorial on autoencoders for nonlinear feature fusion: Taxonomy, models, software and guidelines (2018, methodology): “Denoising autoencoder A denoising AE or DAE [70] learns to generate robust features from inputs by reconstructing partially destroyed samples. The use of AEs for denoising had been introduced earlier [71], but this technique leverages the denoising ability of the AE to build a latent feature space w”
Tagger: Deep Unsupervised Perceptual Grouping (2016, methodology): “Rather than using variational autoencoders, we apply denoising autoencoders [6, 13, 31] which are trained to reconstruct original inputs x from corrupted versions ˜ x .”
Towards the Limit of Network Quantization (2016, methodology): “(cid:12)(cid:12)(cid:12)(cid:12) An efﬁcient way of computing the diagonal of Hessian is presented in Le Cun (1987); Becker & Le Cun (1988) and it is based on the back propagation method that is similar to the back propagation algorithm used for computing ﬁrst-order partial derivatives (gradients).”
Learning to See by Moving (2015, background): “This hypothesis inspired research in fully unsupervised learning from visual imagery and led to popular algorithms such as ICA [5, 20], autoencoders [24, 9], sparse coding [32], restricted boltzmann machines [27, 37] and others [6].”
Neural Networks and Deep Learning (2018, unknown): “More detailed discussions of the autoencoder during its early years were provided in [48, 275].”
Modular Neural Networks and Self-Decomposition (1997, background): “As (Mure et al. 1992) (Jacobs et al. 1991a) (Bottou and Galliinari 1991)highlight, modularity is a way to embed in NN a priori knowledge that is knownto be important to improve NN learning (LeCun 1987).”
Optimal Brain Damage (1989, methodology): “Such a procedure was derived in (Le Cun, 1987), and was the basis of a fast back propagation method used extensively in \1lrious applications (Becker and Le Cun, 1989; Le Cun, 1989; Le Cun et al., 1990a).”
Asynchronous Stochastic Gradient Descent with Delay Compensation (2016, background): “085 0.090 0.095 0.100 0.105 0.110 Test error M = 8 Figure 2. Error rates of the global model w.r.t. number of effective passes of data on CIFAR-10 according to (Choromanska et al.,2015;Kawaguchi,2016;LeCun, 1987)), the condition L 2 >C can be simpliﬁed as L 2 >(1 )L2 1 + L 3ˇ, which is easy to”
Simplifying Neural Networks by Soft Weight-Sharing (1992, background): “A number of authors (e.g., Rumelhart et al. 1986, chapter 8; Lang et al. 1990; le Cun 1989, 1987) have proposed the idea of weight sharing, in which' a single weight is shared among many connections in the network so that the number of adjustable weights in the network is much less than the number…”
Horizontal and Vertical Ensemble with Deep Representation for Classification (2013, background): “An alternative to obtain class prediction is to train a standard classiﬁer (such as Random Forest or SVM) with learned data representations (Bengio et al., 2012).”
A new active labeling method for deep learning (2014, background): “An autoencoder [8-10] is a feedforward NN that attempts to learn an identity function by setting the outputs equal to the inputs (or at least minimizing the reconstruction error), Fig.”
Stochastic Gradient Learning in Neural Networks (1991, methodology): “Cette formulation statistique bien connue a donné lieu à de nombreux résultats théoriques.”
Backpropagation and the brain (2020, background): “Figure 3a sketches target propagation 141,142 , the essential idea behind using a stack of auto-encoders for deep learning.”