ANN - Wiki

10/3/2018 Artificial neural network - Wikipedia
Artificial neural network

Artificial neural networks (ANN) or connectionist
systems are computing systems vaguely inspired by the
biological neural networks that constitute animal brains.[1] Such
systems "learn" to perform tasks by considering examples,
generally without being programmed with any task-specific rules.
For example, in image recognition, they might learn to identify
images that contain cats by analyzing example images that have
been manually labeled as "cat" or "no cat" and using the results to
identify cats in other images. They do this without any prior
knowledge about cats, e.g., that they have fur, tails, whiskers and
cat-like faces. Instead, they automatically generate identifying
characteristics from the learning material that they process.
An ANN is based on a collection of connected units or nodes

called artificial neurons which loosely model the neurons in a
biological brain. Each connection, like the synapses in a
biological brain, can transmit a signal from one artificial neuron
to another. An artificial neuron that receives a signal can process
it and then signal additional artificial neurons connected to it. An artificial neural network is an interconnected
group of nodes, akin to the vast network of
In common ANN implementations, the signal at a connection neurons in a brain. Here, each circular node
between artificial neurons is a real number, and the output of represents an artificial neuron and an arrow
each artificial neuron is computed by some non-linear function of represents a connection from the output of one
the sum of its inputs. The connections between artificial neurons artificial neuron to the input of another.
are called 'edges'. Artificial neurons and edges typically have a
weight that adjusts as learning proceeds. The weight increases or
decreases the strength of the signal at a connection. Artificial neurons may have a threshold such that the signal is only
sent if the aggregate signal crosses that threshold. Typically, artificial neurons are aggregated into layers. Different layers
may perform different kinds of transformations on their inputs. Signals travel from the first layer (the input layer), to the
last layer (the output layer), possibly after traversing the layers multiple times.
The original goal of the ANN approach was to solve problems in the same way that a human brain would. However, over
time, attention moved to performing specific tasks, leading to deviations from biology. Artificial neural networks have
been used on a variety of tasks, including computer vision, speech recognition, machine translation, social network
filtering, playing board and video games and medical diagnosis.
Contents
History
Hebbian learning
Backpropagation
Hardware-based designs
https://en.wikipedia.org/wiki/Artificial_neural_network 1/39
Contests
Convolutional networks
Models
Components of an artificial neural network
Neural networks as functions
Learning
Learning paradigms
Learning algorithms
Variants
Group method of data handling
Convolutional neural networks
Long short-term memory
Deep reservoir computing
Deep belief networks
Large memory storage and retrieval neural networks
Stacked (de-noising) auto-encoders
Deep stacking networks
Tensor deep stacking networks
Spike-and-slab RBMs
Compound hierarchical-deep models
Deep predictive coding networks
Networks with separate memory structures
Multilayer kernel machine
Neural architecture search
Use
Applications
Types of models
Theoretical properties
Computational power
Capacity
Convergence
Generalization and statistics
Criticism
Training issues
Theoretical issues
Hardware issues
Practical counterexamples to criticisms
Hybrid approaches
Types
Gallery
See also
References
Bibliography
External links
History
Warren McCulloch and Walter Pitts[2] (1943) created a computational model for neural networks based on mathematics
and algorithms called threshold logic. This model paved the way for neural network research to split into two approaches.
One approach focused on biological processes in the brain while the other focused on the application of neural networks to
artificial intelligence. This work led to work on nerve networks and their link to finite automata.[3]
Hebbian learning
In the late 1940s, D. O. Hebb[4] created a learning hypothesis based on the mechanism of neural plasticity that became
known as Hebbian learning. Hebbian learning is unsupervised learning. This evolved into models for long term
potentiation. Researchers started applying these ideas to computational models in 1948 with Turing's B-type machines.
Farley and Clark[5] (1954) first used computational machines, then called "calculators", to simulate a Hebbian network.
Other neural network computational machines were created by Rochester, Holland, Habit and Duda (1956).[6]
Rosenblatt[7] (1958) created the perceptron, an algorithm for pattern recognition. With mathematical notation, Rosenblatt
described circuitry not in the basic perceptron, such as the exclusive-or circuit that could not be processed by neural
networks at the time.[8]
In 1959, a biological model proposed by Nobel laureates Hubel and Wiesel was based on their discovery of two types of
cells in the primary visual cortex: simple cells and complex cells.[9]
The first functional networks with many layers were published by Ivakhnenko and Lapa in 1965, becoming the Group
Method of Data Handling.[10][11][12]
Neural network research stagnated after machine learning research by Minsky and Papert (1969),[13] who discovered two
key issues with the computational machines that processed neural networks. The first was that basic perceptrons were
incapable of processing the exclusive-or circuit. The second was that computers didn't have enough processing power to
effectively handle the work required by large neural networks. Neural network research slowed until computers achieved
far greater processing power.
Much of artificial intelligence had focused on high-level (symbolic) models that are processed by using algorithms,
characterized for example by expert systems with knowledge embodied in if-then rules, until in the late 1980s research
expanded to low-level (sub-symbolic) machine learning, characterized by knowledge embodied in the parameters of a
cognitive model.
Backpropagation
A key trigger for renewed interest in neural networks and learning was Werbos's (1975) backpropagation algorithm that
effectively solved the exclusive-or problem by making the training of multi-layer networks feasible and efficient.
Backpropagation distributed the error term back up through the layers, by modifying the weights at each node.[8]
In the mid-1980s, parallel distributed processing became popular under the name connectionism. Rumelhart and
McClelland (1986) described the use of connectionism to simulate neural processes.[14]
Support vector machines and other, much simpler methods such as linear classifiers gradually overtook neural networks
in machine learning popularity. However, using neural networks transformed some domains, such as the prediction of
protein structures.[15][16]
In 1992, max-pooling was introduced to help with least shift invariance and tolerance to deformation to aid in 3D object
recognition.[17][18][19] In 2010, Backpropagation training through max-pooling was accelerated by GPUs and shown to
perform better than other pooling variants.[20]
The vanishing gradient problem affects many-layered feedforward networks that used backpropagation and also recurrent
neural networks (RNNs).[21][22] As errors propagate from layer to layer, they shrink exponentially with the number of
layers, impeding the tuning of neuron weights that is based on those errors, particularly affecting deep networks.
To overcome this problem, Schmidhuber adopted a multi-level hierarchy of networks (1992) pre-trained one level at a
time by unsupervised learning and fine-tuned by backpropagation.[23] Behnke (2003) relied only on the sign of the
gradient (Rprop)[24] on problems such as image reconstruction and face localization.
Hinton et al. (2006) proposed learning a high-level representation using successive layers of binary or real-valued latent
variables with a restricted Boltzmann machine[25] to model each layer. Once sufficiently many layers have been learned,
the deep architecture may be used as a generative model by reproducing the data when sampling down the model (an
"ancestral pass") from the top level feature activations.[26][27] In 2012, Ng and Dean created a network that learned to
recognize higher-level concepts, such as cats, only from watching unlabeled images taken from YouTube videos.[28]
Earlier challenges in training deep neural networks were successfully addressed with methods such as unsupervised pre-
training, while available computing power increased through the use of GPUs and distributed computing. Neural networks
were deployed on a large scale, particularly in image and visual recognition problems. This became known as "deep
learning".
Hardware-based designs
Computational devices were created in CMOS, for both biophysical simulation and neuromorphic computing.
Nanodevices[29] for very large scale principal components analyses and convolution may create a new class of neural
computing because they are fundamentally analog rather than digital (even though the first implementations may use
digital devices).[30] Ciresan and colleagues (2010)[31] in Schmidhuber's group showed that despite the vanishing gradient
problem, GPUs makes back-propagation feasible for many-layered feedforward neural networks.
Contests
Between 2009 and 2012, recurrent neural networks and deep feedforward neural networks developed in Schmidhuber's
research group won eight international competitions in pattern recognition and machine learning.[32][33] For example, the
bi-directional and multi-dimensional long short-term memory (LSTM)[34][35][36][37] of Graves et al. won three competitions
in connected handwriting recognition at the 2009 International Conference on Document Analysis and Recognition
(ICDAR), without any prior knowledge about the three languages to be learned.[36][35]
Ciresan and colleagues won pattern recognition contests, including the IJCNN 2011 Traffic Sign Recognition
Competition,[38] the ISBI 2012 Segmentation of Neuronal Structures in Electron Microscopy Stacks challenge[39] and
others. Their neural networks were the first pattern recognizers to achieve human-competitive or even superhuman
performance[40] on benchmarks such as traffic sign recognition (IJCNN 2012), or the MNIST handwritten digits problem.
Researchers demonstrated (2010) that deep neural networks interfaced to a hidden Markov model with context-
dependent states that define the neural network output layer can drastically reduce errors in large-vocabulary speech
recognition tasks such as voice search.
GPU-based implementations[41] of this approach won many pattern recognition contests, including the IJCNN 2011
Traffic Sign Recognition Competition,[38] the ISBI 2012 Segmentation of neuronal structures in EM stacks challenge,[39]
the ImageNet Competition[42] and others.
Deep, highly nonlinear neural architectures similar to the neocognitron[43] and the "standard architecture of vision",[44]
inspired by simple and complex cells, were pre-trained by unsupervised methods by Hinton.[45][26] A team from his lab
won a 2012 contest sponsored by Merck to design software to help find molecules that might identify new drugs.[46]
Convolutional networks
As of 2011, the state of the art in deep learning feedforward networks alternated between convolutional layers and max-
pooling layers,[41][47] topped by several fully or sparsely connected layers followed by a final classification layer. Learning
is usually done without unsupervised pre-training.
Such supervised deep learning methods were the first to achieve human-competitive performance on certain tasks.[40]
Artificial neural networks were able to guarantee shift invariance to deal with small and large natural objects in large
cluttered scenes, only when invariance extended beyond shift, to all ANN-learned concepts, such as location, type (object
class label), scale, lighting and others. This was realized in Developmental Networks (DNs)[48] whose embodiments are
Where-What Networks, WWN-1 (2008)[49] through WWN-7 (2013).[50]
Models
An artificial neural network is a network of
simple elements called artificial neurons, which
receive input, change their internal state
(activation) according to that input, and produce
output depending on the input and activation.
An artificial neuron mimics the working of a

biophysical neuron with inputs and outputs, but is
not a biological neuron model.
The network forms by connecting the output of

certain neurons to the input of other neurons Neuron and myelinated axon, with signal flow from inputs at
forming a directed, weighted graph. The weights dendrites to outputs at axon terminals
as well as the functions that compute the
activation can be modified by a process called
learning which is governed by a learning rule.[51]
Components of an artificial neural network
Neurons
A neuron with label receiving an input from predecessor neurons consists of the following components:[51]
an activation , depending on a discrete time parameter,

possibly a threshold , which stays fixed unless changed by a learning function,
an activation function that computes the new activation at a given time from , and the net input
giving rise to the relation
and an output function computing the output from the activation
Often the output function is simply the Identity function.
An input neuron has no predecessor but serves as input interface for the whole network. Similarly an output neuron has
no successor and thus serves as output interface of the whole network.
Connections, weights and biases

The network consists of connections, each connection transferring the output of a neuron to the input of a neuron . In
this sense is the predecessor of and is the successor of . Each connection is assigned a weight .[51] Sometimes a
bias term added to total weighted sum of inputs to serve as threshold to shift the activation function.[52]
Propagation function
The propagation function computes the input to the neuron from the outputs of predecessor neurons and
typically has the form[51]
When a bias value added with the function, the above form changes to following [53]
, where is a bias.
Learning rule
The learning rule is a rule or an algorithm which modifies the parameters of the neural network, in order for a given input
to the network to produce a favored output. This learning process typically amounts to modifying the weights and
thresholds of the variables within the network.[51]
Neural networks as functions

Neural network models can be viewed as simple mathematical models defining a function or a distribution
over or both and . Sometimes models are intimately associated with a particular learning rule. A common use of
the phrase "ANN model" is really the definition of a class of such functions (where members of the class are obtained by
varying parameters, connection weights, or specifics of the architecture such as the number of neurons or their
connectivity).
Mathematically, a neuron's network function is defined as a composition of other functions , that can further be
decomposed into other functions. This can be conveniently represented as a network structure, with arrows depicting the
dependencies between functions. A widely used type of composition is the nonlinear weighted sum, where
, where (commonly referred to as the activation function[54]) is some predefined function, such
as the hyperbolic tangent or sigmoid function or softmax function or rectifier function. The important characteristic of the
activation function is that it provides a smooth transition as input values change, i.e. a small change in input produces a
small change in output. The following refers to a collection of functions as a vector .
This figure depicts such a decomposition of , with dependencies between variables

indicated by arrows. These can be interpreted in two ways.
The first view is the functional view: the input is transformed into a 3-dimensional vector
, which is then transformed into a 2-dimensional vector , which is finally transformed
into . This view is most commonly encountered in the context of optimization.
ANN dependency graph
The second view is the probabilistic view: the random variable depends upon
the random variable , which depends upon , which depends upon the
random variable . This view is most commonly encountered in the context of graphical models.
The two views are largely equivalent. In either case, for this particular architecture, the components of individual layers
are independent of each other (e.g., the components of are independent of each other given their input ). This naturally
enables a degree of parallelism in the implementation.
Networks such as the previous one are commonly called feedforward, because their graph is a
directed acyclic graph. Networks with cycles are commonly called recurrent. Such networks are
commonly depicted in the manner shown at the top of the figure, where is shown as being
dependent upon itself. However, an implied temporal dependence is not shown.
Learning
The possibility of learning has attracted the most interest in neural networks. Given a specific Two separate
task to solve, and a class of functions , learning means using a set of observations to find depictions of the
which solves the task in some optimal sense. recurrent ANN
dependency graph
This entails defining a cost function such that, for the optimal solution ,
– i.e., no solution has a cost less than the cost of the optimal solution
(see mathematical optimization).
The cost function is an important concept in learning, as it is a measure of how far away a particular solution is from an
optimal solution to the problem to be solved. Learning algorithms search through the solution space to find a function that
has the smallest possible cost.
For applications where the solution is data dependent, the cost must necessarily be a function of the observations,
otherwise the model would not relate to the data. It is frequently defined as a statistic to which only approximations can be
made. As a simple example, consider the problem of finding the model , which minimizes , for data
pairs drawn from some distribution . In practical situations we would only have samples from and thus, for
the above example, we would only minimize . Thus, the cost is minimized over a sample of the
data rather than the entire distribution.
When some form of online machine learning must be used, where the cost is reduced as each new example is
seen. While online machine learning is often used when is fixed, it is most useful in the case where the distribution
changes slowly over time. In neural network methods, some form of online machine learning is frequently used for finite
datasets.
Choosing a cost function

While it is possible to define an ad hoc cost function, frequently a particular cost (function) is used, either because it has
desirable properties (such as convexity) or because it arises naturally from a particular formulation of the problem (e.g., in
a probabilistic formulation the posterior probability of the model can be used as an inverse cost). Ultimately, the cost
function depends on the task.
Backpropagation
A DNN can be discriminatively trained with the standard backpropagation algorithm. Backpropagation is a method to
calculate the gradient of the loss function (produces the cost associated with a given state) with respect to the weights in
an ANN.
The basics of continuous backpropagation[10][55][56][57] were derived in the context of control theory by Kelley[58] in 1960
and by Bryson in 1961,[59] using principles of dynamic programming. In 1962, Dreyfus published a simpler derivation
based only on the chain rule.[60] Bryson and Ho described it as a multi-stage dynamic system optimization method in
1969.[61][62] In 1970, Linnainmaa finally published the general method for automatic differentiation (AD) of discrete
connected networks of nested differentiable functions.[63][64] This corresponds to the modern version of backpropagation
which is efficient even when the networks are sparse.[10][55][65][66] In 1973, Dreyfus used backpropagation to adapt
parameters of controllers in proportion to error gradients.[67] In 1974, Werbos mentioned the possibility of applying this
principle to Artificial neural networks,[68] and in 1982, he applied Linnainmaa's AD method to neural networks in the way
that is widely used today.[55][69] In 1986, Rumelhart, Hinton and Williams noted that this method can generate useful
internal representations of incoming data in hidden layers of neural networks.[70] In 1993, Wan was the first[10] to win an
international pattern recognition contest through backpropagation.[71]
The weight updates of backpropagation can be done via stochastic gradient descent using the following equation:
where, is the learning rate, is the cost (loss) function and a stochastic term. The choice of the cost function
depends on factors such as the learning type (supervised, unsupervised, reinforcement, etc.) and the activation function.
For example, when performing supervised learning on a multiclass classification problem, common choices for the
activation function and cost function are the softmax function and cross entropy function, respectively. The softmax
function is defined as where represents the class probability (output of the unit ) and and
represent the total input to units and of the same level respectively. Cross entropy is defined as
where represents the target probability for output unit and is the probability output for after applying the
activation function.[72]
These can be used to output object bounding boxes in the form of a binary mask. They are also used for multi-scale
regression to increase localization precision. DNN-based regression can learn features that capture geometric information
in addition to serving as a good classifier. They remove the requirement to explicitly model parts and their relations. This
helps to broaden the variety of objects that can be learned. The model consists of multiple layers, each of which has a
rectified linear unit as its activation function for non-linear transformation. Some layers are convolutional, while others
are fully connected. Every convolutional layer has an additional max pooling. The network is trained to minimize L2 error
for predicting the mask ranging over the entire training set containing bounding boxes represented as masks.
Alternatives to backpropagation include Extreme Learning Machines,[73] "No-prop" networks,[74] training without
backtracking,[75] "weightless" networks,[76][77] and non-connectionist neural networks.
Learning paradigms
The three major learning paradigms each correspond to a particular learning task. These are supervised learning,
unsupervised learning and reinforcement learning.
Supervised learning
Supervised learning uses a set of example pairs and the aim is to find a function in the
allowed class of functions that matches the examples. In other words, we wish to infer the mapping implied by the data;
the cost function is related to the mismatch between our mapping and the data and it implicitly contains prior knowledge
about the problem domain.[78]
A commonly used cost is the mean-squared error, which tries to minimize the average squared error between the
network's output, , and the target value over all the example pairs. Minimizing this cost using gradient descent for
the class of neural networks called multilayer perceptrons (MLP), produces the backpropagation algorithm for training
neural networks.
Tasks that fall within the paradigm of supervised learning are pattern recognition (also known as classification) and
regression (also known as function approximation). The supervised learning paradigm is also applicable to sequential data
(e.g., for hand writing, speech and gesture recognition). This can be thought of as learning with a "teacher", in the form of
a function that provides continuous feedback on the quality of solutions obtained thus far.
Unsupervised learning
In unsupervised learning, some data is given and the cost function to be minimized, that can be any function of the data
and the network's output, .
The cost function is dependent on the task (the model domain) and any a priori assumptions (the implicit properties of
the model, its parameters and the observed variables).
As a trivial example, consider the model where is a constant and the cost . Minimizing
this cost produces a value of that is equal to the mean of the data. The cost function can be much more complicated. Its
form depends on the application: for example, in compression it could be related to the mutual information between and
, whereas in statistical modeling, it could be related to the posterior probability of the model given the data (note that
in both of those examples those quantities would be maximized rather than minimized).
Tasks that fall within the paradigm of unsupervised learning are in general estimation problems; the applications include
clustering, the estimation of statistical distributions, compression and filtering.
Reinforcement learning
In reinforcement learning, data are usually not given, but generated by an agent's interactions with the environment. At
each point in time , the agent performs an action and the environment generates an observation and an
instantaneous cost , according to some (usually unknown) dynamics. The aim is to discover a policy for selecting actions
that minimizes some measure of a long-term cost, e.g., the expected cumulative cost. The environment's dynamics and the
long-term cost for each policy are usually unknown, but can be estimated.
More formally the environment is modeled as a Markov decision process (MDP) with states and actions
with the following probability distributions: the instantaneous cost distribution , the observation
distribution and the transition , while a policy is defined as the conditional distribution over
actions given the observations. Taken together, the two then define a Markov chain (MC). The aim is to discover the policy
(i.e., the MC) that minimizes the cost.
Artificial neural networks are frequently used in reinforcement learning as part of the overall algorithm.[79][80] Dynamic
programming was coupled with Artificial neural networks (giving neurodynamic programming) by Bertsekas and
Tsitsiklis[81] and applied to multi-dimensional nonlinear problems such as those involved in vehicle routing,[82] natural
resources management[83][84] or medicine[85] because of the ability of Artificial neural networks to mitigate losses of
accuracy even when reducing the discretization grid density for numerically approximating the solution of the original
control problems.
Tasks that fall within the paradigm of reinforcement learning are control problems, games and other sequential decision
making tasks.
Learning algorithms
Training a neural network model essentially means selecting one model from the set of allowed models (or, in a Bayesian
framework, determining a distribution over the set of allowed models) that minimizes the cost. Numerous algorithms are
available for training neural network models; most of them can be viewed as a straightforward application of optimization
theory and statistical estimation.
Most employ some form of gradient descent, using backpropagation to compute the actual gradients. This is done by
simply taking the derivative of the cost function with respect to the network parameters and then changing those
parameters in a gradient-related direction. Backpropagation training algorithms fall into three categories:
steepest descent (with variable learning rate and momentum, resilient backpropagation);
quasi-Newton (Broyden-Fletcher-Goldfarb-Shanno, one step secant);
Levenberg-Marquardt and conjugate gradient (Fletcher-Reeves update, Polak-Ribiére update, Powell-Beale restart,
scaled conjugate gradient).[86]
Evolutionary methods,[87] gene expression programming,[88] simulated annealing,[89] expectation-maximization, non-
parametric methods and particle swarm optimization[90] are other methods for training neural networks.
Convergent recursive learning algorithm

This is a learning method specially designed for cerebellar model articulation controller (CMAC) neural networks. In
2004, a recursive least squares algorithm was introduced to train CMAC neural network online.[91] This algorithm can
converge in one step and update all weights in one step with any new input data. Initially, this algorithm had
computational complexity of O(N3). Based on QR decomposition, this recursive learning algorithm was simplified to be
O(N).[92]
Variants

The Group Method of Data Handling (GMDH)[93] features fully automatic structural and parametric model optimization.
The node activation functions are Kolmogorov-Gabor polynomials that permit additions and multiplications. It used a
deep feedforward multilayer perceptron with eight layers.[94] It is a supervised learning network that grows layer by layer,
where each layer is trained by regression analysis. Useless items are detected using a validation set, and pruned through
regularization. The size and depth of the resulting network depends on the task.[95]
Convolutional neural networks

A convolutional neural network (CNN) is a class of deep, feed-forward networks, composed of one or more convolutional
layers with fully connected layers (matching those in typical Artificial neural networks) on top. It uses tied weights and
pooling layers. In particular, max-pooling[18] is often structured via Fukushima's convolutional architecture.[96] This
architecture allows CNNs to take advantage of the 2D structure of input data.
CNNs are suitable for processing visual and other two-dimensional data.[97][98] They have shown superior results in both
image and speech applications. They can be trained with standard backpropagation. CNNs are easier to train than other
regular, deep, feed-forward neural networks and have many fewer parameters to estimate.[99] Examples of applications in
computer vision include DeepDream[100] and robot navigation.[101]
A recent development has been that of Capsule Neural Network (CapsNet), the idea behind which is to add structures
called capsules to a CNN and to reuse output from several of those capsules to form more stable (with respect to various
perturbations) representations for higher order capsules.[102]
Long short-term memory

Long short-term memory (LSTM) networks are RNNs that avoid the vanishing gradient problem.[103] LSTM is normally
augmented by recurrent gates called forget gates.[104] LSTM networks prevent backpropagated errors from vanishing or
exploding.[21] Instead errors can flow backwards through unlimited numbers of virtual layers in space-unfolded LSTM.
That is, LSTM can learn "very deep learning" tasks[10] that require memories of events that happened thousands or even
millions of discrete time steps ago. Problem-specific LSTM-like topologies can be evolved.[105] LSTM can handle long
delays and signals that have a mix of low and high frequency components.
Stacks of LSTM RNNs[106] trained by Connectionist Temporal Classification (CTC)[107] can find an RNN weight matrix
that maximizes the probability of the label sequences in a training set, given the corresponding input sequences. CTC
achieves both alignment and recognition.
In 2003, LSTM started to become competitive with traditional speech recognizers.[108] In 2007, the combination with CTC
achieved first good results on speech data.[109] In 2009, a CTC-trained LSTM was the first RNN to win pattern recognition
contests, when it won several competitions in connected handwriting recognition.[10][36] In 2014, Baidu used CTC-trained
RNNs to break the Switchboard Hub5'00 speech recognition benchmark, without traditional speech processing
methods.[110] LSTM also improved large-vocabulary speech recognition,[111][112] text-to-speech synthesis,[113] for Google
Android,[55][114] and photo-real talking heads.[115] In 2015, Google's speech recognition experienced a 49% improvement
through CTC-trained LSTM.[116]
LSTM became popular in Natural Language Processing. Unlike previous models based on HMMs and similar concepts,
LSTM can learn to recognise context-sensitive languages.[117] LSTM improved machine translation,[118][119] language
modeling[120] and multilingual language processing.[121] LSTM combined with CNNs improved automatic image
captioning.[122]
Deep reservoir computing

Deep Reservoir Computing and Deep Echo State Networks (deepESNs)[123][124] provide a framework for efficiently trained
models for hierarchical processing of temporal data, while enabling the investigation of the inherent role of RNN layered
composition.
Deep belief networks

A deep belief network (DBN) is a probabilistic, generative model made up of
multiple layers of hidden units. It can be considered a composition of simple
learning modules that make up each layer.[125]
A DBN can be used to generatively pre-train a DNN by using the learned DBN
weights as the initial DNN weights. Backpropagation or other discriminative
algorithms can then tune these weights. This is particularly helpful when
training data are limited, because poorly initialized weights can significantly
hinder model performance. These pre-trained weights are in a region of the
weight space that is closer to the optimal weights than were they randomly
chosen. This allows for both improved modeling and faster convergence of the
fine-tuning phase.[126]
A restricted Boltzmann machine
(RBM) with fully connected visible
Large memory storage and retrieval neural networks and hidden units. Note there are no
hidden-hidden or visible-visible
Large memory storage and retrieval neural networks (LAMSTAR)[127][128] are
connections.
fast deep learning neural networks of many layers that can use many filters
simultaneously. These filters may be nonlinear, stochastic, logic, non-
stationary, or even non-analytical. They are biologically motivated and learn continuously.
A LAMSTAR neural network may serve as a dynamic neural network in spatial or time domains or both. Its speed is
provided by Hebbian link-weights[129] that integrate the various and usually different filters (preprocessing functions) into
its many layers and to dynamically rank the significance of the various layers and functions relative to a given learning
task. This grossly imitates biological learning which integrates various preprocessors (cochlea, retina, etc.) and cortexes
(auditory, visual, etc.) and their various regions. Its deep learning capability is further enhanced by using inhibition,
correlation and its ability to cope with incomplete data, or "lost" neurons or layers even amidst a task. It is fully
transparent due to its link weights. The link-weights allow dynamic determination of innovation and redundancy, and
facilitate the ranking of layers, of filters or of individual neurons relative to a task.
LAMSTAR has been applied to many domains, including medical[130][131][132] and financial predictions,[133] adaptive
filtering of noisy speech in unknown noise,[134] still-image recognition,[135] video image recognition,[136] software
security[137] and adaptive control of non-linear systems.[138] LAMSTAR had a much faster learning speed and somewhat
lower error rate than a CNN based on ReLU-function filters and max pooling, in 20 comparative studies.[139]
These applications demonstrate delving into aspects of the data that are hidden from shallow learning networks and the
human senses, such as in the cases of predicting onset of sleep apnea events,[131] of an electrocardiogram of a fetus as
recorded from skin-surface electrodes placed on the mother's abdomen early in pregnancy,[132] of financial prediction[127]
or in blind filtering of noisy speech.[134]
LAMSTAR was proposed in 1996 (A U.S. Patent 5,920,852 A (https://www.google.com/patents/US5920852)) and was
further developed Graupe and Kordylewski from 1997–2002.[140][141][142] A modified version, known as LAMSTAR 2, was
developed by Schneider and Graupe in 2008.[143][144]
Stacked (de-noising) auto-encoders

The auto encoder idea is motivated by the concept of a good representation. For example, for a classifier, a good
representation can be defined as one that yields a better-performing classifier.
An encoder is a deterministic mapping that transforms an input vector x into hidden representation y, where
, is the weight matrix and b is an offset vector (bias). A decoder maps back the hidden representation y to
the reconstructed input z via . The whole process of auto encoding is to compare this reconstructed input to the original
and try to minimize the error to make the reconstructed value as close as possible to the original.
In stacked denoising auto encoders, the partially corrupted output is cleaned (de-noised). This idea was introduced in
2010 by Vincent et al.[145] with a specific approach to good representation, a good representation is one that can be
obtained robustly from a corrupted input and that will be useful for recovering the corresponding clean input. Implicit in
this definition are the following ideas:
The higher level representations are relatively stable and robust to input corruption;
It is necessary to extract features that are useful for representation of the input distribution.
The algorithm starts by a stochastic mapping of to through , this is the corrupting step. Then the corrupted
input passes through a basic auto-encoder process and is mapped to a hidden representation .
From this hidden representation, we can reconstruct . In the last stage, a minimization algorithm runs in order
to have z as close as possible to uncorrupted input . The reconstruction error might be either the cross-entropy
loss with an affine-sigmoid decoder, or the squared error loss with an affine decoder.[145]
In order to make a deep architecture, auto encoders stack.[146] Once the encoding function of the first denoising auto
encoder is learned and used to uncorrupt the input (corrupted input), the second level can be trained.[145]
Once the stacked auto encoder is trained, its output can be used as the input to a supervised learning algorithm such as
support vector machine classifier or a multi-class logistic regression.[145]
Deep stacking networks

A deep stacking network (DSN)[147] (deep convex network) is based on a hierarchy of blocks of simplified neural network
modules. It was introduced in 2011 by Deng and Dong.[148] It formulates the learning as a convex optimization problem
with a closed-form solution, emphasizing the mechanism's similarity to stacked generalization.[149] Each DSN block is a
simple module that is easy to train by itself in a supervised fashion without backpropagation for the entire blocks.[150]
Each block consists of a simplified multi-layer perceptron (MLP) with a single hidden layer. The hidden layer h has
logistic sigmoidal units, and the output layer has linear units. Connections between these layers are represented by weight
matrix U; input-to-hidden-layer connections have weight matrix W. Target vectors t form the columns of matrix T, and
the input data vectors x form the columns of matrix X. The matrix of hidden units is . Modules are
trained in order, so lower-layer weights W are known at each stage. The function performs the element-wise logistic
sigmoid operation. Each block estimates the same final label class y, and its estimate is concatenated with original input X
to form the expanded input for the next block. Thus, the input to the first block contains the original data only, while
downstream blocks' input adds the output of preceding blocks. Then learning the upper-layer weight matrix U given other
weights in the network can be formulated as a convex optimization problem:
which has a closed-form solution.
Unlike other deep architectures, such as DBNs, the goal is not to discover the transformed feature representation. The
structure of the hierarchy of this kind of architecture makes parallel learning straightforward, as a batch-mode
optimization problem. In purely discriminative tasks, DSNs perform better than conventional DBNs.[147]
Tensor deep stacking networks

This architecture is a DSN extension. It offers two important improvements: it uses higher-order information from
covariance statistics, and it transforms the non-convex problem of a lower-layer to a convex sub-problem of an upper-
layer.[151] TDSNs use covariance statistics in a bilinear mapping from each of two distinct sets of hidden units in the same
layer to predictions, via a third-order tensor.
While parallelization and scalability are not considered seriously in conventional DNNs,[152][153][154] all learning for DSNs
and TDSNs is done in batch mode, to allow parallelization.[148][147] Parallelization allows scaling the design to larger
(deeper) architectures and data sets.
The basic architecture is suitable for diverse tasks such as classification and regression.
Spike-and-slab RBMs
The need for deep learning with real-valued inputs, as in Gaussian restricted Boltzmann machines, led to the spike-and-
slab RBM (ssRBM), which models continuous-valued inputs with strictly binary latent variables.[155] Similar to basic
RBMs and its variants, a spike-and-slab RBM is a bipartite graph, while like GRBMs, the visible units (input) are real-
valued. The difference is in the hidden layer, where each hidden unit has a binary spike variable and a real-valued slab
variable. A spike is a discrete probability mass at zero, while a slab is a density over continuous domain;[156] their mixture
forms a prior.[157]
An extension of ssRBM called µ-ssRBM provides extra modeling capacity using additional terms in the energy function.
One of these terms enables the model to form a conditional distribution of the spike variables by marginalizing out the
slab variables given an observation.
Compound hierarchical-deep models

Compound hierarchical-deep models compose deep networks with non-parametric Bayesian models. Features can be
learned using deep architectures such as DBNs,[26] DBMs,[158] deep auto encoders,[159] convolutional variants,[160][161]
ssRBMs,[156] deep coding networks,[162] DBNs with sparse feature learning,[163] RNNs,[164] conditional DBNs,[165] de-
noising auto encoders.[166] This provides a better representation, allowing faster learning and more accurate classification
with high-dimensional data. However, these architectures are poor at learning novel classes with few examples, because
all network units are involved in representing the input (a distributed representation) and must be adjusted together
(high degree of freedom). Limiting the degree of freedom reduces the number of parameters to learn, facilitating learning
of new classes from few examples. Hierarchical Bayesian (HB) models allow learning from few examples, for
example[167][168][169][170][171] for computer vision, statistics and cognitive science.
Compound HD architectures aim to integrate characteristics of both HB and deep networks. The compound HDP-DBM
architecture is a hierarchical Dirichlet process (HDP) as a hierarchical model, incorporated with DBM architecture. It is a
full generative model, generalized from abstract concepts flowing through the layers of the model, which is able to
synthesize new examples in novel classes that look "reasonably" natural. All the levels are learned jointly by maximizing a
joint log-probability score.[172]
In a DBM with three hidden layers, the probability of a visible input ν is:
where is the set of hidden units, and are the model parameters,
representing visible-hidden and hidden-hidden symmetric interaction terms.
A learned DBM model is an undirected model that defines the joint distribution . One way to express what
has been learned is the conditional model and a prior term .
Here represents a conditional DBM model, which can be viewed as a two-layer DBM but with bias terms
given by the states of :
Deep predictive coding networks

A deep predictive coding network (DPCN) is a predictive coding scheme that uses top-down information to empirically
adjust the priors needed for a bottom-up inference procedure by means of a deep, locally connected, generative model.
This works by extracting sparse features from time-varying observations using a linear dynamical model. Then, a pooling
strategy is used to learn invariant feature representations. These units compose to form a deep architecture and are
trained by greedy layer-wise unsupervised learning. The layers constitute a kind of Markov chain such that the states at
any layer depend only on the preceding and succeeding layers.
DPCNs predict the representation of the layer, by using a top-down approach using the information in upper layer and
temporal dependencies from previous states.[173]
DPCNs can be extended to form a convolutional network.[173]
Networks with separate memory structures

Integrating external memory with Artificial neural networks dates to early research in distributed representations[174] and
Kohonen's self-organizing maps. For example, in sparse distributed memory or hierarchical temporal memory, the
patterns encoded by neural networks are used as addresses for content-addressable memory, with "neurons" essentially
serving as address encoders and decoders. However, the early controllers of such memories were not differentiable.
LSTM-related differentiable memory structures

Apart from long short-term memory (LSTM), other approaches also added differentiable memory to recurrent functions.
For example:
Differentiable push and pop actions for alternative memory networks called neural stack machines[175][176]
Memory networks where the control network's external differentiable storage is in the fast weights of another
network[177]
LSTM forget gates[178]
Self-referential RNNs with special output units for addressing and rapidly manipulating the RNN's own weights in
differentiable fashion (internal storage)[179][180]
Learning to transduce with unbounded memory[181]
Neural Turing machines

Neural Turing machines[182] couple LSTM networks to external memory resources, with which they can interact by
attentional processes. The combined system is analogous to a Turing machine but is differentiable end-to-end, allowing it
to be efficiently trained by gradient descent. Preliminary results demonstrate that neural Turing machines can infer simple
algorithms such as copying, sorting and associative recall from input and output examples.
Differentiable neural computers (DNC) are an NTM extension. They out-performed Neural turing machines, long short-
term memory systems and memory networks on sequence-processing tasks.[183][184][185][186][187]
Semantic hashing
Approaches that represent previous experiences directly and use a similar experience to form a local model are often
called nearest neighbour or k-nearest neighbors methods.[188] Deep learning is useful in semantic hashing[189] where a
deep graphical model the word-count vectors[190] obtained from a large set of documents. Documents are mapped to
memory addresses in such a way that semantically similar documents are located at nearby addresses. Documents similar
to a query document can then be found by accessing all the addresses that differ by only a few bits from the address of the
query document. Unlike sparse distributed memory that operates on 1000-bit addresses, semantic hashing works on 32 or
64-bit addresses found in a conventional computer architecture.
Memory networks
Memory networks[191][192] are another extension to neural networks incorporating long-term memory. The long-term
memory can be read and written to, with the goal of using it for prediction. These models have been applied in the context
of question answering (QA) where the long-term memory effectively acts as a (dynamic) knowledge base and the output is
a textual response.[193] A team of electrical and computer engineers from UCLA Samueli School of Engineering has created
a physical artificial neural network. That can analyze large volumes of data and identify objects at the actual speed of
light.[194]
Pointer networks
Deep neural networks can be potentially improved by deepening and parameter reduction, while maintaining trainability.
While training extremely deep (e.g., 1 million layers) neural networks might not be practical, CPU-like architectures such
as pointer networks[195] and neural random-access machines[196] overcome this limitation by using external random-
access memory and other components that typically belong to a computer architecture such as registers, ALU and
pointers. Such systems operate on probability distribution vectors stored in memory cells and registers. Thus, the model is
fully differentiable and trains end-to-end. The key characteristic of these models is that their depth, the size of their short-
term memory, and the number of parameters can be altered independently – unlike models like LSTM, whose number of
parameters grows quadratically with memory size.
Encoder–decoder networks
Encoder–decoder frameworks are based on neural networks that map highly structured input to highly structured output.
The approach arose in the context of machine translation,[197][198][199] where the input and output are written sentences in
two natural languages. In that work, an LSTM RNN or CNN was used as an encoder to summarize a source sentence, and
the summary was decoded using a conditional RNN language model to produce the translation.[200] These systems share
building blocks: gated RNNs and CNNs and trained attention mechanisms.
Multilayer kernel machine

Multilayer kernel machines (MKM) are a way of learning highly nonlinear functions by iterative application of weakly
nonlinear kernels. They use the kernel principal component analysis (KPCA),[201] as a method for the unsupervised greedy
layer-wise pre-training step of deep learning.[202]
Layer learns the representation of the previous layer , extracting the principal component (PC) of the projection
layer output in the feature domain induced by the kernel. For the sake of dimensionality reduction of the updated
representation in each layer, a supervised strategy selects the best informative features among features extracted by KPCA.
The process is:
rank the features according to their mutual information with the class labels;
for different values of K and , compute the classification error rate of a K-nearest neighbor (K-NN)
classifier using only the most informative features on a validation set;
the value of with which the classifier has reached the lowest error rate determines the number of features to
retain.
Some drawbacks accompany the KPCA method as the building cells of an MKM.
A more straightforward way to use kernel machines for deep learning was developed for spoken language
understanding.[203] The main idea is to use a kernel machine to approximate a shallow neural net with an infinite number
of hidden units, then use stacking to splice the output of the kernel machine and the raw input in building the next, higher
level of the kernel machine. The number of levels in the deep convex network is a hyper-parameter of the overall system,
to be determined by cross validation.
Neural architecture search

Neural architecture search (NAS) uses machine learning to automate the design of Artificial neural networks. Various
approaches to NAS have designed networks that compare well with hand-designed systems. The basic search algorithm is
to propose a candidate model, evaluate it against a dataset and use the results as feedback to teach the NAS network.[204]
Use
Using Artificial neural networks requires an understanding of their characteristics.
Choice of model: This depends on the data representation and the application. Overly complex models slow learning.
Learning algorithm: Numerous trade-offs exist between learning algorithms. Almost any algorithm will work well with
the correct hyperparameters for training on a particular data set. However, selecting and tuning an algorithm for
training on unseen data requires significant experimentation.
Robustness: If the model, cost function and learning algorithm are selected appropriately, the resulting ANN can
become robust.
ANN capabilities fall within the following broad categories:
Function approximation, or regression analysis, including time series prediction, fitness approximation and modeling.
Classification, including pattern and sequence recognition, novelty detection and sequential decision making.
Data processing, including filtering, clustering, blind source separation and compression.
Robotics, including directing manipulators and prostheses.
Control, including computer numerical control.
Applications
Because of their ability to reproduce and model nonlinear processes, Artificial neural networks have found many
applications in a wide range of disciplines.
Application areas include system identification and control (vehicle control, trajectory prediction,[205] process control,
natural resource management), quantum chemistry,[206] game-playing and decision making (backgammon, chess, poker),
pattern recognition (radar systems, face identification, signal classification,[207] object recognition and more), sequence
recognition (gesture, speech, handwritten and printed text recognition), medical diagnosis, finance[208] (e.g. automated
trading systems), data mining, visualization, machine translation, social network filtering[209] and e-mail spam filtering.
Artificial neural networks have been used to diagnose cancers, including lung cancer,[210] prostate cancer, colorectal
cancer[211] and to distinguish highly invasive cancer cell lines from less invasive lines using only cell shape
information.[212][213]
Artificial neural networks have been used to accelerate reliability analysis of infrastructures subject to natural
disasters.[214][215]
Artificial neural networks have also been used for building black-box models in geoscience: hydrology,[216][217] ocean
modelling and coastal engineering,[218][219] and geomorphology.[220]
Types of models
Many types of models are used, defined at different levels of abstraction and modeling different aspects of neural systems.
They range from models of the short-term behavior of individual neurons,[221] models of how the dynamics of neural
circuitry arise from interactions between individual neurons and finally to models of how behavior can arise from abstract
neural modules that represent complete subsystems. These include models of the long-term, and short-term plasticity, of
neural systems and their relations to learning and memory from the individual neuron to the system level.
Theoretical properties
Computational power
The multilayer perceptron is a universal function approximator, as proven by the universal approximation theorem.
However, the proof is not constructive regarding the number of neurons required, the network topology, the weights and
the learning parameters.
A specific recurrent architecture with rational valued weights (as opposed to full precision real number-valued weights)
has the full power of a universal Turing machine,[222] using a finite number of neurons and standard linear connections.
Further, the use of irrational values for weights results in a machine with super-Turing power.[223]
Capacity
Models' "capacity" property roughly corresponds to their ability to model any given function. It is related to the amount of
information that can be stored in the network and to the notion of complexity.
Convergence
Models may not consistently converge on a single solution, firstly because many local minima may exist, depending on the
cost function and the model. Secondly, the optimization method used might not guarantee to converge when it begins far
from any local minimum. Thirdly, for sufficiently large data or parameters, some methods become impractical. However,
for CMAC neural network, a recursive least squares algorithm was introduced to train it, and this algorithm can be
guaranteed to converge in one step.[91]
Generalization and statistics

Applications whose goal is to create a system that generalizes well to unseen examples, face the possibility of over-training.
This arises in convoluted or over-specified systems when the capacity of the network significantly exceeds the needed free
parameters. Two approaches address over-training. The first is to use cross-validation and similar techniques to check for
the presence of over-training and optimally select hyperparameters to minimize the generalization error. The second is to
use some form of regularization. This concept emerges in a probabilistic (Bayesian) framework, where regularization can
be performed by selecting a larger prior probability over simpler models; but also in statistical learning theory, where the
goal is to minimize over two quantities: the 'empirical risk' and the 'structural risk', which roughly corresponds to the error
over the training set and the predicted error in unseen data due to overfitting.
Supervised neural networks that use a mean squared error (MSE) cost function
can use formal statistical methods to determine the confidence of the trained
model. The MSE on a validation set can be used as an estimate for variance.
This value can then be used to calculate the confidence interval of the output of
the network, assuming a normal distribution. A confidence analysis made this
way is statistically valid as long as the output probability distribution stays the
same and the network is not modified.
By assigning a softmax activation function, a generalization of the logistic

function, on the output layer of the neural network (or a softmax component in Confidence analysis of a neural
a component-based neural network) for categorical target variables, the network
outputs can be interpreted as posterior probabilities. This is very useful in
classification as it gives a certainty measure on classifications.
The softmax activation function is:
Criticism
Training issues
A common criticism of neural networks, particularly in robotics, is that they require too much training for real-world
operation. Potential solutions include randomly shuffling training examples, by using a numerical optimization algorithm
that does not take too large steps when changing the network connections following an example and by grouping examples
in so-called mini-batches. Improving the training efficiency and convergence capability has always been an ongoing
research area for neural network. For example, by introducing a recursive least squares algorithm for CMAC neural
network, the training process only takes one step to converge.[91]
Theoretical issues
No neural network has solved computationally difficult problems such as the n-Queens problem, the travelling salesman
problem, or the problem of factoring large integers.
A fundamental objection is that they do not reflect how real neurons function. Back propagation is a critical part of most
artificial neural networks, although no such mechanism exists in biological neural networks.[224] How information is
coded by real neurons is not known. Sensor neurons fire action potentials more frequently with sensor activation and
muscle cells pull more strongly when their associated motor neurons receive action potentials more frequently.[225] Other
than the case of relaying information from a sensor neuron to a motor neuron, almost nothing of the principles of how
information is handled by biological neural networks is known.
The motivation behind Artificial neural networks is not necessarily to strictly replicate neural function, but to use
biological neural networks as an inspiration. A central claim of Artificial neural networks is therefore that it embodies
some new and powerful general principle for processing information. Unfortunately, these general principles are ill-
defined. It is often claimed that they are emergent from the network itself. This allows simple statistical association (the
basic function of artificial neural networks) to be described as learning or recognition. Alexander Dewdney commented
that, as a result, artificial neural networks have a "something-for-nothing quality, one that imparts a peculiar aura of
laziness and a distinct lack of curiosity about just how good these computing systems are. No human hand (or mind)
intervenes; solutions are found as if by magic; and no one, it seems, has learned anything".[226]
Biological brains use both shallow and deep circuits as reported by brain anatomy,[227] displaying a wide variety of
invariance. Weng[228] argued that the brain self-wires largely according to signal statistics and therefore, a serial cascade
cannot catch all major statistical dependencies.
Hardware issues
Large and effective neural networks require considerable computing resources.[229] While the brain has hardware tailored
to the task of processing signals through a graph of neurons, simulating even a simplified neuron on von Neumann
architecture may compel a neural network designer to fill many millions of database rows for its connections – which can
consume vast amounts of memory and storage. Furthermore, the designer often needs to transmit signals through many
of these connections and their associated neurons – which must often be matched with enormous CPU processing power
and time.
Schmidhuber notes that the resurgence of neural networks in the twenty-first century is largely attributable to advances in
hardware: from 1991 to 2015, computing power, especially as delivered by GPGPUs (on GPUs), has increased around a
million-fold, making the standard backpropagation algorithm feasible for training networks that are several layers deeper
than before.[230] The use of parallel GPUs can reduce training times from months to days.[229]
Neuromorphic engineering addresses the hardware difficulty directly, by constructing non-von-Neumann chips to directly
implement neural networks in circuitry. Another chip optimized for neural network processing is called a Tensor
Processing Unit, or TPU.[231]
Practical counterexamples to criticisms

Arguments against Dewdney's position are that neural networks have been successfully used to solve many complex and
diverse tasks, ranging from autonomously flying aircraft[232] to detecting credit card fraud to mastering the game of Go.
Technology writer Roger Bridgman commented:
Neural networks, for instance, are in the dock not only because they have been hyped to high heaven, (what
hasn't?) but also because you could create a successful net without understanding how it worked: the bunch
of numbers that captures its behaviour would in all probability be "an opaque, unreadable table...valueless
as a scientific resource".
In spite of his emphatic declaration that science is not technology, Dewdney seems here to pillory neural
nets as bad science when most of those devising them are just trying to be good engineers. An unreadable
table that a useful machine could read would still be well worth having.[233]
Although it is true that analyzing what has been learned by an artificial neural network is difficult, it is much easier to do
so than to analyze what has been learned by a biological neural network. Furthermore, researchers involved in exploring
learning algorithms for neural networks are gradually uncovering general principles that allow a learning machine to be
successful. For example, local vs non-local learning and shallow vs deep architecture.[234]
Hybrid approaches
Advocates of hybrid models (combining neural networks and symbolic approaches), claim that such a mixture can better
capture the mechanisms of the human mind.[235][236]
Types
Artificial neural networks have many variations. The simplest, static types have one or more static components, including
number of units, number of layers, unit weights and topology. Dynamic types allow one or more of these to change during
the learning process. The latter are much more complicated, but can shorten learning periods and produce better results.
Some types allow/require learning to be "supervised" by the operator, while others operate independently. Some types
operate purely in hardware, while others are purely software and run on general purpose computers.
Gallery
A single-layer feedforward artificial neural A two-layer feedforward artificial neural network.

network. Arrows originating from are omitted
for clarity. There are p inputs to this network and
q outputs. In this system, the value of the qth
output, would be calculated as
An artificial neural network. An ANN dependency graph.
A single-layer feedforward artificial neural A two-layer feedforward artificial neural network

network with 4 inputs, 6 hidden and 2 outputs. with 8 inputs, 2x8 hidden and 2 outputs. Given
Given position state and direction outputs wheel position state, direction and other environment
based control values. values outputs thruster based control values.
Parallel pipeline structure of CMAC neural

network. This learning algorithm can converge in
one step.
See also
Hierarchical temporal memory
20Q
ADALINE
Adaptive resonance theory
Artificial life
Associative memory
Autoencoder
BEAM robotics
Biological cybernetics
Biologically inspired computing
Blue Brain Project
Catastrophic interference
Cerebellar Model Articulation Controller (CMAC)
Cognitive architecture
Cognitive science
Convolutional neural network (CNN)
Connectionist expert system
Connectomics
Cultured neuronal networks
Deep learning
Encog
Fuzzy logic
Gene expression programming
Genetic algorithm
Genetic programming
Habituation
In Situ Adaptive Tabulation
Machine learning concepts
Models of neural computation
Neuroevolution
Neural coding
Neural gas
Neural machine translation
Neural network software
Neuroscience
Nonlinear system identification
Optical neural network
Parallel Constraint Satisfaction Processes
Parallel distributed processing
Radial basis function network
Recurrent neural networks
Self-organizing map
Spiking neural network
Systolic array
Tensor product network
Time delay neural network (TDNN)
References
1. "Artificial Neural Networks as Models of Neural Information Processing | Frontiers Research Topic" (https://www.fronti
ersin.org/research-topics/4817/artificial-neural-networks-as-models-of-neural-information-processing). Retrieved
2018-02-20.
2. McCulloch, Warren; Walter Pitts (1943). "A Logical Calculus of Ideas Immanent in Nervous Activity". Bulletin of
Mathematical Biophysics. 5 (4): 115–133. doi:10.1007/BF02478259 (https://doi.org/10.1007%2FBF02478259).
3. Kleene, S.C. (1956). "Representation of Events in Nerve Nets and Finite Automata" (https://www.degruyter.com/view/
books/9781400882618/9781400882618-002/9781400882618-002.xml). Annals of Mathematics Studies (34).
Princeton University Press. pp. 3–41. Retrieved 2017-06-17.
4. Hebb, Donald (1949). The Organization of Behavior (https://books.google.com/books?id=ddB4AgAAQBAJ). New
York: Wiley. ISBN 978-1-135-63190-1.
5. Farley, B.G.; W.A. Clark (1954). "Simulation of Self-Organizing Systems by Digital Computer". IRE Transactions on
Information Theory. 4 (4): 76–84. doi:10.1109/TIT.1954.1057468 (https://doi.org/10.1109%2FTIT.1954.1057468).
6. Rochester, N.; J.H. Holland; L.H. Habit; W.L. Duda (1956). "Tests on a cell assembly theory of the action of the brain,
using a large digital computer". IRE Transactions on Information Theory. 2 (3): 80–93. doi:10.1109/TIT.1956.1056810
(https://doi.org/10.1109%2FTIT.1956.1056810).
7. Rosenblatt, F. (1958). "The Perceptron: A Probabilistic Model For Information Storage And Organization In The
Brain". Psychological Review. 65 (6): 386–408. CiteSeerX 10.1.1.588.3775 (https://citeseerx.ist.psu.edu/viewdoc/sum
mary?doi=10.1.1.588.3775). doi:10.1037/h0042519 (https://doi.org/10.1037%2Fh0042519). PMID 13602029 (https://
www.ncbi.nlm.nih.gov/pubmed/13602029).
8. Werbos, P.J. (1975). Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences (https://b
ooks.google.com/books?id=z81XmgEACAAJ).
9. David H. Hubel and Torsten N. Wiesel (2005). Brain and visual perception: the story of a 25-year collaboration (http
s://books.google.com/books?id=8YrxWojxUA4C&pg=PA106). Oxford University Press US. p. 106. ISBN 978-0-19-
517618-6.
10. Schmidhuber, J. (2015). "Deep Learning in Neural Networks: An Overview". Neural Networks. 61: 85–117.
arXiv:1404.7828 (https://arxiv.org/abs/1404.7828). doi:10.1016/j.neunet.2014.09.003 (https://doi.org/10.1016%2Fj.ne
unet.2014.09.003). PMID 25462637 (https://www.ncbi.nlm.nih.gov/pubmed/25462637).
11. Ivakhnenko, A. G. (1973). Cybernetic Predicting Devices (https://books.google.com/books?id=FhwVNQAACAAJ).
CCM Information Corporation.
12. Ivakhnenko, A. G.; Grigorʹevich Lapa, Valentin (1967). Cybernetics and forecasting techniques (https://books.google.c
om/books?id=rGFgAAAAMAAJ). American Elsevier Pub. Co.
13. Minsky, Marvin; Papert, Seymour (1969). Perceptrons: An Introduction to Computational Geometry (https://books.goo
gle.com/books?id=Ow1OAQAAIAAJ). MIT Press. ISBN 0-262-63022-2.
14. Rumelhart, D.E; McClelland, James (1986). Parallel Distributed Processing: Explorations in the Microstructure of
Cognition (https://books.google.com/books?id=davmLgzusB8C). Cambridge: MIT Press. ISBN 978-0-262-63110-5.
15. Qian, N.; Sejnowski, T.J. (1988). "Predicting the secondary structure of globular proteins using neural network
models". Journal of Molecular Biology. 202. pp. 865–884. Qian1988.
16. Rost, B.; Sander, C. (1993). "Prediction of protein secondary structure at better than 70% accuracy". Journal of
Molecular Biology. 232. pp. 584–599. Rost1993.
17. J. Weng, N. Ahuja and T. S. Huang, "Cresceptron: a self-organizing neural network which grows adaptively (http://ww
w.cse.msu.edu/~weng/research/CresceptronIJCNN1992.pdf)," Proc. International Joint Conference on Neural
Networks, Baltimore, Maryland, vol I, pp. 576–581, June, 1992.
18. J. Weng, N. Ahuja and T. S. Huang, "Learning recognition and segmentation of 3-D objects from 2-D images (http://w
ww.cse.msu.edu/~weng/research/CresceptronICCV1993.pdf)," Proc. 4th International Conf. Computer Vision, Berlin,
Germany, pp. 121–128, May, 1993.
19. J. Weng, N. Ahuja and T. S. Huang, "Learning recognition and segmentation using the Cresceptron (http://www.cse.m
su.edu/~weng/research/CresceptronIJCV.pdf)," International Journal of Computer Vision, vol. 25, no. 2, pp. 105–139,
Nov. 1997.
20. Dominik Scherer, Andreas C. Müller, and Sven Behnke: "Evaluation of Pooling Operations in Convolutional
Architectures for Object Recognition (https://www.ais.uni-bonn.de/papers/icann2010_maxpool.pdf)," In 20th
International Conference Artificial Neural Networks (ICANN), pp. 92–101, 2010. doi:10.1007/978-3-642-15825-4_10
(https://doi.org/10.1007%2F978-3-642-15825-4_10).
21. S. Hochreiter., "Untersuchungen zu dynamischen neuronalen Netzen (http://people.idsia.ch/~juergen/SeppHochreiter
1991ThesisAdvisorSchmidhuber.pdf)," Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J.
Schmidhuber, 1991.
22. Hochreiter, S.; et al. (15 January 2001). "Gradient flow in recurrent nets: the difficulty of learning long-term
dependencies". In Kolen, John F.; Kremer, Stefan C. A Field Guide to Dynamical Recurrent Networks (https://books.g
oogle.com/books?id=NWOcMVA64aAC). John Wiley & Sons. ISBN 978-0-7803-5369-5.
23. J. Schmidhuber., "Learning complex, extended sequences using the principle of history compression," Neural
Computation, 4, pp. 234–242, 1992.
24. Sven Behnke (2003). Hierarchical Neural Networks for Image Interpretation (http://www.ais.uni-bonn.de/books/LNCS2
766.pdf) (PDF). Lecture Notes in Computer Science. 2766. Springer.
25. Smolensky, P. (1986). "Information processing in dynamical systems: Foundations of harmony theory.". In D. E.
Rumelhart, J. L. McClelland, & the PDP Research Group. Parallel Distributed Processing: Explorations in the
Microstructure of Cognition (http://portal.acm.org/citation.cfm?id=104290). 1. pp. 194–281.
26. Hinton, G. E.; Osindero, S.; Teh, Y. (2006). "A fast learning algorithm for deep belief nets" (http://www.cs.toronto.edu/~
hinton/absps/fastnc.pdf) (PDF). Neural Computation. 18 (7): 1527–1554. doi:10.1162/neco.2006.18.7.1527 (https://do
i.org/10.1162%2Fneco.2006.18.7.1527). PMID 16764513 (https://www.ncbi.nlm.nih.gov/pubmed/16764513).
27. Hinton, G. (2009). "Deep belief networks" (http://www.scholarpedia.org/article/Deep_belief_networks). Scholarpedia.
4 (5): 5947. Bibcode:2009SchpJ...4.5947H (http://adsabs.harvard.edu/abs/2009SchpJ...4.5947H).
doi:10.4249/scholarpedia.5947 (https://doi.org/10.4249%2Fscholarpedia.5947).
28. Ng, Andrew; Dean, Jeff (2012). "Building High-level Features Using Large Scale Unsupervised Learning".
arXiv:1112.6209 (https://arxiv.org/abs/1112.6209) [cs.LG (https://arxiv.org/archive/cs.LG)].
29. Yang, J. J.; Pickett, M. D.; Li, X. M.; Ohlberg, D. A. A.; Stewart, D. R.; Williams, R. S. (2008). "Memristive switching
mechanism for metal/oxide/metal nanodevices". Nat. Nanotechnol. 3 (7): 429–433. doi:10.1038/nnano.2008.160 (http
s://doi.org/10.1038%2Fnnano.2008.160).
30. Strukov, D. B.; Snider, G. S.; Stewart, D. R.; Williams, R. S. (2008). "The missing memristor found". Nature. 453
(7191): 80–83. Bibcode:2008Natur.453...80S (http://adsabs.harvard.edu/abs/2008Natur.453...80S).
doi:10.1038/nature06932 (https://doi.org/10.1038%2Fnature06932). PMID 18451858 (https://www.ncbi.nlm.nih.gov/p
ubmed/18451858).
31. Cireşan, Dan Claudiu; Meier, Ueli; Gambardella, Luca Maria; Schmidhuber, Jürgen (2010-09-21). "Deep, Big, Simple
Neural Nets for Handwritten Digit Recognition" (http://www.mitpressjournals.org/doi/10.1162/NECO_a_00052). Neural
Computation. 22 (12): 3207–3220. doi:10.1162/neco_a_00052 (https://doi.org/10.1162%2Fneco_a_00052).
ISSN 0899-7667 (https://www.worldcat.org/issn/0899-7667).
32. 2012 Kurzweil AI Interview (http://www.kurzweilai.net/how-bio-inspired-deep-learning-keeps-winning-competitions)
with Jürgen Schmidhuber on the eight competitions won by his Deep Learning team 2009–2012
33. "How bio-inspired deep learning keeps winning competitions | KurzweilAI" (http://www.kurzweilai.net/how-bio-inspired-
deep-learning-keeps-winning-competitions). www.kurzweilai.net. Retrieved 2017-06-16.
34. Graves, Alex; and Schmidhuber, Jürgen; Offline Handwriting Recognition with Multidimensional Recurrent Neural
Networks (http://www.idsia.ch/~juergen/nips2009.pdf), in Bengio, Yoshua; Schuurmans, Dale; Lafferty, John; Williams,
Chris K. I.; and Culotta, Aron (eds.), Advances in Neural Information Processing Systems 22 (NIPS'22), 7–10
December 2009, Vancouver, BC, Neural Information Processing Systems (NIPS) Foundation, 2009, pp. 545–552.
35. Graves, A.; Liwicki, M.; Fernandez, S.; Bertolami, R.; Bunke, H.; Schmidhuber, J. (2009). "A Novel Connectionist
System for Improved Unconstrained Handwriting Recognition" (http://www.idsia.ch/~juergen/tpami_2008.pdf) (PDF).
IEEE Transactions on Pattern Analysis and Machine Intelligence. 31 (5): 855–868. doi:10.1109/tpami.2008.137 (http
s://doi.org/10.1109%2Ftpami.2008.137).
36. Graves, Alex; Schmidhuber, Jürgen (2009). Bengio, Yoshua; Schuurmans, Dale; Lafferty, John; Williams, Chris editor-
K. I.; Culotta, Aron, eds. "Offline Handwriting Recognition with Multidimensional Recurrent Neural Networks" (https://p
apers.nips.cc/paper/3449-offline-handwriting-recognition-with-multidimensional-recurrent-neural-networks). Neural
Information Processing Systems (NIPS) Foundation: 545–552.
37. Graves, A.; Liwicki, M.; Fernández, S.; Bertolami, R.; Bunke, H.; Schmidhuber, J. (May 2009). "A Novel Connectionist
System for Unconstrained Handwriting Recognition" (http://ieeexplore.ieee.org/document/4531750/). IEEE
Transactions on Pattern Analysis and Machine Intelligence. 31 (5): 855–868. doi:10.1109/tpami.2008.137 (https://doi.
org/10.1109%2Ftpami.2008.137). ISSN 0162-8828 (https://www.worldcat.org/issn/0162-8828).
38. Cireşan, Dan; Meier, Ueli; Masci, Jonathan; Schmidhuber, Jürgen (August 2012). "Multi-column deep neural network
for traffic sign classification" (http://www.sciencedirect.com/science/article/pii/S0893608012000524). Neural Networks.
Selected Papers from IJCNN 2011. 32: 333–338. doi:10.1016/j.neunet.2012.02.023 (https://doi.org/10.1016%2Fj.neu
net.2012.02.023).
39. Ciresan, Dan; Giusti, Alessandro; Gambardella, Luca M.; Schmidhuber, Juergen (2012). Pereira, F.; Burges, C. J. C.;
Bottou, L.; Weinberger, K. Q., eds. Advances in Neural Information Processing Systems 25 (http://papers.nips.cc/pap
er/4741-deep-neural-networks-segment-neuronal-membranes-in-electron-microscopy-images.pdf) (PDF). Curran
Associates, Inc. pp. 2843–2851.
40. Ciresan, Dan; Meier, U.; Schmidhuber, J. (June 2012). "Multi-column deep neural networks for image classification" (h
ttp://ieeexplore.ieee.org/document/6248110/). 2012 IEEE Conference on Computer Vision and Pattern Recognition:
3642–3649. arXiv:1202.2745 (https://arxiv.org/abs/1202.2745). doi:10.1109/cvpr.2012.6248110 (https://doi.org/10.110
9%2Fcvpr.2012.6248110). ISBN 978-1-4673-1228-8.
41. Ciresan, D. C.; Meier, U.; Masci, J.; Gambardella, L. M.; Schmidhuber, J. (2011). "Flexible, High Performance
Convolutional Neural Networks for Image Classification" (http://ijcai.org/papers11/Papers/IJCAI11-210.pdf) (PDF).
International Joint Conference on Artificial Intelligence. doi:10.5591/978-1-57735-516-8/ijcai11-210 (https://doi.org/10.
5591%2F978-1-57735-516-8%2Fijcai11-210).
42. Krizhevsky, Alex; Sutskever, Ilya; Hinton, Geoffry (2012). "ImageNet Classification with Deep Convolutional Neural
Networks" (https://www.cs.toronto.edu/~kriz/imagenet_classification_with_deep_convolutional.pdf) (PDF). NIPS 2012:
Neural Information Processing Systems, Lake Tahoe, Nevada.
43. Fukushima, K. (1980). "Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition
unaffected by shift in position". Biological Cybernetics. 36 (4): 93–202. doi:10.1007/BF00344251 (https://doi.org/10.10
07%2FBF00344251). PMID 7370364 (https://www.ncbi.nlm.nih.gov/pubmed/7370364).
44. Riesenhuber, M; Poggio, T (1999). "Hierarchical models of object recognition in cortex". Nature Neuroscience. 2 (11):
1019–1025. doi:10.1038/14819 (https://doi.org/10.1038%2F14819).
45. Hinton, Geoffrey (2009-05-31). "Deep belief networks" (http://www.scholarpedia.org/article/Deep_belief_networks).
Scholarpedia. 4 (5): 5947. Bibcode:2009SchpJ...4.5947H (http://adsabs.harvard.edu/abs/2009SchpJ...4.5947H).
doi:10.4249/scholarpedia.5947 (https://doi.org/10.4249%2Fscholarpedia.5947). ISSN 1941-6016 (https://www.worldc
at.org/issn/1941-6016).
46. Markoff, John (November 23, 2012). "Scientists See Promise in Deep-Learning Programs" (https://www.nytimes.com/
2012/11/24/science/scientists-see-advances-in-deep-learning-a-part-of-artificial-intelligence.html). New York Times.
47. Martines, H.; Bengio, Y.; Yannakakis, G. N. (2013). "Learning Deep Physiological Models of Affect". IEEE
Computational Intelligence. 8 (2): 20–33. doi:10.1109/mci.2013.2247823 (https://doi.org/10.1109%2Fmci.2013.22478
23).
48. J. Weng, "Why Have We Passed 'Neural Networks Do not Abstract Well'? (http://www.cse.msu.edu/~weng/research/
WhyPass-Weng-NI-2011.pdf)," Natural Intelligence: the INNS Magazine, vol. 1, no.1, pp. 13–22, 2011.
49. Z. Ji, J. Weng, and D. Prokhorov, "Where-What Network 1: Where and What Assist Each Other Through Top-down
Connections (http://www.cse.msu.edu/~weng/research/ICDL08_0077.pdf)," Proc. 7th International Conference on
Development and Learning (ICDL'08), Monterey, CA, Aug. 9–12, pp. 1–6, 2008.
50. X. Wu, G. Guo, and J. Weng, "Skull-closed Autonomous Development: WWN-7 Dealing with Scales (http://www.cse.
msu.edu/~weng/research/WWN7-Wu-ICBM-2013.pdf)," Proc. International Conference on Brain-Mind, July 27–28,
East Lansing, Michigan, pp. 1–9, 2013.
51. Zell, Andreas (1994). "chapter 5.2". Simulation Neuronaler Netze [Simulation of Neural Networks] (in German) (1st
ed.). Addison-Wesley. ISBN 3-89319-554-8.
52. Abbod, Maysam F (2007). "Application of Artificial Intelligence to the Management of Urological Cancer" (https://www.
sciencedirect.com/science/article/pii/S0022534707013936). The Journal of Urology. 178 (4): 1150–1156.
doi:10.1016/j.juro.2007.05.122 (https://doi.org/10.1016%2Fj.juro.2007.05.122).
53. DAWSON, CHRISTIAN W (1998). "An artificial neural network approach to rainfall-runoff modelling" (https://www.tand
fonline.com/doi/abs/10.1080/02626669809492102). Hydrological Sciences Journal. 43 (1): 47–66.
doi:10.1080/02626669809492102 (https://doi.org/10.1080%2F02626669809492102).
54. "The Machine Learning Dictionary" (http://www.cse.unsw.edu.au/~billw/mldict.html#activnfn).
55. Schmidhuber, Jürgen (2015). "Deep Learning" (http://www.scholarpedia.org/article/Deep_Learning). Scholarpedia. 10
(11): 32832. Bibcode:2015SchpJ..1032832S (http://adsabs.harvard.edu/abs/2015SchpJ..1032832S).
doi:10.4249/scholarpedia.32832 (https://doi.org/10.4249%2Fscholarpedia.32832).
56. Dreyfus, Stuart E. (1990-09-01). "Artificial neural networks, back propagation, and the Kelley-Bryson gradient
procedure" (http://arc.aiaa.org/doi/10.2514/3.25422). Journal of Guidance, Control, and Dynamics. 13 (5): 926–928.
Bibcode:1990JGCD...13..926D (http://adsabs.harvard.edu/abs/1990JGCD...13..926D). doi:10.2514/3.25422 (https://d
oi.org/10.2514%2F3.25422). ISSN 0731-5090 (https://www.worldcat.org/issn/0731-5090).
57. Eiji Mizutani, Stuart Dreyfus, Kenichi Nishio (2000). On derivation of MLP backpropagation from the Kelley-Bryson
optimal-control gradient formula and its application. Proceedings of the IEEE International Joint Conference on Neural
Networks (IJCNN 2000), Como Italy, July 2000. Online (http://queue.ieor.berkeley.edu/People/Faculty/dreyfus-pubs/ijc
nn2k.pdf)
58. Kelley, Henry J. (1960). "Gradient theory of optimal flight paths" (http://arc.aiaa.org/doi/abs/10.2514/8.5282?journalCo
de=arsj). ARS Journal. 30 (10): 947–954. doi:10.2514/8.5282 (https://doi.org/10.2514%2F8.5282).
59. Arthur E. Bryson (1961, April). A gradient method for optimizing multi-stage allocation processes. In Proceedings of
the Harvard Univ. Symposium on digital computers and their applications.
60. Dreyfus, Stuart (1962). "The numerical solution of variational problems" (https://www.researchgate.net/publication/256
244271_The_numerical_solution_of_variational_problems). Journal of Mathematical Analysis and Applications. 5 (1):
30–45. doi:10.1016/0022-247x(62)90004-5 (https://doi.org/10.1016%2F0022-247x%2862%2990004-5).
61. Russell, Stuart J.; Norvig, Peter (2010). Artificial Intelligence A Modern Approach (https://books.google.com/books?id
=8jZBksh-bUMC&pg=PA578). Prentice Hall. p. 578. ISBN 978-0-13-604259-4. "The most popular method for learning
in multilayer networks is called Back-propagation."
62. Bryson, Arthur Earl (1969). Applied Optimal Control: Optimization, Estimation and Control (https://books.google.com/b
ooks?id=1bChDAEACAAJ&pg=PA481). Blaisdell Publishing Company or Xerox College Publishing. p. 481.
63. Seppo Linnainmaa (1970). The representation of the cumulative rounding error of an algorithm as a Taylor expansion
of the local rounding errors. Master's Thesis (in Finnish), Univ. Helsinki, 6–7.
64. Linnainmaa, Seppo (1976). "Taylor expansion of the accumulated rounding error". BIT Numerical Mathematics. 16
(2): 146–160. doi:10.1007/bf01931367 (https://doi.org/10.1007%2Fbf01931367).
65. Griewank, Andreas (2012). "Who Invented the Reverse Mode of Differentiation?" (https://web.archive.org/web/201707
21211929/http://www.math.uiuc.edu/documenta/vol-ismp/52_griewank-andreas-b.pdf) (PDF). Documenta
Matematica, Extra Volume ISMP: 389–400. Archived from the original (http://www.math.uiuc.edu/documenta/vol-ismp/
52_griewank-andreas-b.pdf) (PDF) on 21 July 2017. Retrieved 27 June 2017.
66. Griewank, Andreas; Walther, Andrea (2008). Evaluating Derivatives: Principles and Techniques of Algorithmic
Differentiation, Second Edition (https://books.google.com/books?id=xoiiLaRxcbEC). SIAM. ISBN 978-0-89871-776-1.
67. Dreyfus, Stuart (1973). "The computational solution of optimal control problems with time lag". IEEE Transactions on
Automatic Control. 18 (4): 383–385. doi:10.1109/tac.1973.1100330 (https://doi.org/10.1109%2Ftac.1973.1100330).
68. Paul Werbos (1974). Beyond regression: New tools for prediction and analysis in the behavioral sciences. PhD thesis,
Harvard University.
69. Werbos, Paul (1982). "Applications of advances in nonlinear sensitivity analysis". System modeling and optimization
(http://werbos.com/Neural/SensitivityIFIPSeptember1981.pdf) (PDF). Springer. pp. 762–770.
70. Rumelhart, David E.; Hinton, Geoffrey E.; Williams, Ronald J. (1986). "Learning representations by back-propagating
errors" (http://www.nature.com/articles/Art323533a0). Nature. 323 (6088): 533–536. Bibcode:1986Natur.323..533R (ht
tp://adsabs.harvard.edu/abs/1986Natur.323..533R). doi:10.1038/323533a0 (https://doi.org/10.1038%2F323533a0).
71. Eric A. Wan (1993). "Time series prediction by using a connectionist network with internal delay lines." In Proceedings
of the Santa Fe Institute Studies in the Sciences of Complexity, 15: p. 195. Addison-Wesley Publishing Co.
72. Hinton, G.; Deng, L.; Yu, D.; Dahl, G. E.; Mohamed, A. r; Jaitly, N.; Senior, A.; Vanhoucke, V.; Nguyen, P. (November
2012). "Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research
Groups" (http://ieeexplore.ieee.org/document/6296526/). IEEE Signal Processing Magazine. 29 (6): 82–97.
Bibcode:2012ISPM...29...82H (http://adsabs.harvard.edu/abs/2012ISPM...29...82H). doi:10.1109/msp.2012.2205597
(https://doi.org/10.1109%2Fmsp.2012.2205597). ISSN 1053-5888 (https://www.worldcat.org/issn/1053-5888).
73. Huang, Guang-Bin; Zhu, Qin-Yu; Siew, Chee-Kheong (2006). "Extreme learning machine: theory and applications".
Neurocomputing. 70 (1): 489–501. doi:10.1016/j.neucom.2005.12.126 (https://doi.org/10.1016%2Fj.neucom.2005.12.
126).
74. Widrow, Bernard; et al. (2013). "The no-prop algorithm: A new learning algorithm for multilayer neural networks".
Neural Networks. 37: 182–188. doi:10.1016/j.neunet.2012.09.020 (https://doi.org/10.1016%2Fj.neunet.2012.09.020).
75. Ollivier, Yann; Charpiat, Guillaume (2015). "Training recurrent networks without backtracking". arXiv:1507.07680 (http
s://arxiv.org/abs/1507.07680) [cs.NE (https://arxiv.org/archive/cs.NE)].
76. ESANN. 2009
77. Hinton, G. E. (2010). "A Practical Guide to Training Restricted Boltzmann Machines" (https://www.researchgate.net/pu
blication/221166159_A_brief_introduction_to_Weightless_Neural_Systems). Tech. Rep. UTML TR 2010-003,.
78. Ojha, Varun Kumar; Abraham, Ajith; Snášel, Václav (2017-04-01). "Metaheuristic design of feedforward neural
networks: A review of two decades of research" (http://www.sciencedirect.com/science/article/pii/S095219761730023
4). Engineering Applications of Artificial Intelligence. 60: 97–116. doi:10.1016/j.engappai.2017.01.013 (https://doi.org/
10.1016%2Fj.engappai.2017.01.013).
79. Dominic, S.; Das, R.; Whitley, D.; Anderson, C. (July 1991). "Genetic reinforcement learning for neural networks".
IJCNN-91-Seattle International Joint Conference on Neural Networks. IJCNN-91-Seattle International Joint
Conference on Neural Networks. Seattle, Washington, USA: IEEE. doi:10.1109/IJCNN.1991.155315 (https://doi.org/1
0.1109%2FIJCNN.1991.155315). ISBN 0-7803-0164-1.
80. Hoskins, J.C.; Himmelblau, D.M. (1992). "Process control via artificial neural networks and reinforcement learning".
Computers & Chemical Engineering. 16 (4): 241–251. doi:10.1016/0098-1354(92)80045-B (https://doi.org/10.1016%2
F0098-1354%2892%2980045-B).
81. Bertsekas, D.P.; Tsitsiklis, J.N. (1996). Neuro-dynamic programming (https://papers.nips.cc/paper/4741-deep-neural-n
etworks-segment-neuronal-membranes-in-electron-microscopy-images). Athena Scientific. p. 512. ISBN 1-886529-
10-8.
82. Secomandi, Nicola (2000). "Comparing neuro-dynamic programming algorithms for the vehicle routing problem with
stochastic demands". Computers & Operations Research. 27 (11–12): 1201–1225. doi:10.1016/S0305-
0548(99)00146-X (https://doi.org/10.1016%2FS0305-0548%2899%2900146-X).
83. de Rigo, D.; Rizzoli, A. E.; Soncini-Sessa, R.; Weber, E.; Zenesi, P. (2001). "Neuro-dynamic programming for the
efficient management of reservoir networks" (https://zenodo.org/record/7482/files/de_Rigo_etal_MODSIM2001_active
link_authorcopy.pdf) (PDF). Proceedings of MODSIM 2001, International Congress on Modelling and Simulation.
MODSIM 2001, International Congress on Modelling and Simulation (http://www.mssanz.org.au/MODSIM01/MODSIM
01.htm). Canberra, Australia: Modelling and Simulation Society of Australia and New Zealand.
doi:10.5281/zenodo.7481 (https://doi.org/10.5281%2Fzenodo.7481). ISBN 0-867405252. Retrieved 29 July 2012.
84. Damas, M.; Salmeron, M.; Diaz, A.; Ortega, J.; Prieto, A.; Olivares, G. (2000). "Genetic algorithms and neuro-
dynamic programming: application to water supply networks". Proceedings of 2000 Congress on Evolutionary
Computation. 2000 Congress on Evolutionary Computation. La Jolla, California, USA: IEEE.
doi:10.1109/CEC.2000.870269 (https://doi.org/10.1109%2FCEC.2000.870269). ISBN 0-7803-6375-2.
85. Deng, Geng; Ferris, M.C. (2008). "Neuro-dynamic programming for fractionated radiotherapy planning". Springer
Optimization and Its Applications. Springer Optimization and Its Applications. 12: 47–70. CiteSeerX 10.1.1.137.8288
(https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.137.8288). doi:10.1007/978-0-387-73299-2_3 (https://doi.
org/10.1007%2F978-0-387-73299-2_3). ISBN 978-0-387-73298-5.
86. M. Forouzanfar; H. R. Dajani; V. Z. Groza; M. Bolic & S. Rajan (July 2010). Comparison of Feed-Forward Neural
Network Training Algorithms for Oscillometric Blood Pressure Estimation (https://www.researchgate.net/profile/Moha
mad_Forouzanfar/publication/224173336_Comparison_of_Feed-Forward_Neural_Network_training_algorithms_for_o
scillometric_blood_pressure_estimation/links/00b7d533829c3a7484000000.pdf?ev=pub_int_doc_dl&origin=publicatio
n_detail&inViewer=true&msrp=TyT96%2BjWOHJo%2BVhkMF4IzwHPAImSd442n%2BAkEuXj9qBmQSZ495CpxqlaO
Yon%2BSlEzWQElBGyJmbBCiiUOV8ImeEqPFXiIRivcrWsWmlPBYU%3D) (PDF). 4th Int. Workshop Soft Computing
Applications. Arad, Romania: IEEE.
87. de Rigo, D.; Castelletti, A.; Rizzoli, A. E.; Soncini-Sessa, R.; Weber, E. (January 2005). "A selective improvement
technique for fastening Neuro-Dynamic Programming in Water Resources Network Management" (http://www.nt.ntnu.
no/users/skoge/prost/proceedings/ifac2005/Papers/Paper4269.html). In Pavel Zítek. Proceedings of the 16th IFAC
World Congress – IFAC-PapersOnLine. 16th IFAC World Congress (http://www.nt.ntnu.no/users/skoge/prost/proceedi
ngs/ifac2005/Index.html). 16. Prague, Czech Republic: IFAC. doi:10.3182/20050703-6-CZ-1902.02172 (https://doi.or
g/10.3182%2F20050703-6-CZ-1902.02172). ISBN 978-3-902661-75-3. Retrieved 30 December 2011.
88. Ferreira, C. (2006). "Designing Neural Networks Using Gene Expression Programming" (http://www.gene-expression-
programming.com/webpapers/Ferreira-ASCT2006.pdf) (PDF). In A. Abraham, B. de Baets, M. Köppen, and B.
Nickolay, eds., Applied Soft Computing Technologies: The Challenge of Complexity, pages 517–536, Springer-Verlag.
89. Da, Y.; Xiurun, G. (July 2005). T. Villmann, ed. An improved PSO-based ANN with simulated annealing technique.
New Aspects in Neurocomputing: 11th European Symposium on Artificial Neural Networks (http://www.dice.ucl.ac.be/
esann/proceedings/electronicproceedings.htm). Elsevier. doi:10.1016/j.neucom.2004.07.002 (https://doi.org/10.101
6%2Fj.neucom.2004.07.002).
90. Wu, J.; Chen, E. (May 2009). Wang, H., Shen, Y., Huang, T., Zeng, Z., eds. A Novel Nonparametric Regression
Ensemble for Rainfall Forecasting Using Particle Swarm Optimization Technique Coupled with Artificial Neural
Network. 6th International Symposium on Neural Networks, ISNN 2009 (http://www2.mae.cuhk.edu.hk/~isnn2009/).
Springer. doi:10.1007/978-3-642-01513-7-6 (https://doi.org/10.1007%2F978-3-642-01513-7-6). ISBN 978-3-642-
01215-0.
91. Ting Qin, et al. "A learning algorithm of CMAC based on RLS." Neural Processing Letters 19.1 (2004): 49–61.
92. Ting Qin, et al. "Continuous CMAC-QRLS and its systolic array." Neural Processing Letters 22.1 (2005): 1–16.
93. Ivakhnenko, Alexey Grigorevich (1968). "The group method of data handling – a rival of the method of stochastic
approximation". Soviet Automatic Control. 13 (3): 43–55.
94. Ivakhnenko, Alexey (1971). "Polynomial theory of complex systems". IEEE Transactions on Systems, Man and
Cybernetics (4) (4): 364–378. doi:10.1109/TSMC.1971.4308320 (https://doi.org/10.1109%2FTSMC.1971.4308320).
95. Kondo, T.; Ueno, J. (2008). "Multi-layered GMDH-type neural network self-selecting optimum neural network
architecture and its application to 3-dimensional medical image recognition of blood vessels" (https://www.researchga
te.net/publication/228402366_GMDH-Type_Neural_Network_Self-Selecting_Optimum_Neural_Network_Architecture
_and_Its_Application_to_3-Dimensional_Medical_Image_Recognition_of_the_Lungs). International Journal of
Innovative Computing, Information and Control. 4 (1): 175–187.
96. Fukushima, K. (1980). "Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition
unaffected by shift in position". Biol. Cybern. 36 (4): 193–202. doi:10.1007/bf00344251 (https://doi.org/10.1007%2Fbf
00344251). PMID 7370364 (https://www.ncbi.nlm.nih.gov/pubmed/7370364).
97. LeCun et al., "Backpropagation Applied to Handwritten Zip Code Recognition," Neural Computation, 1, pp. 541–551,
1989.
98. Yann LeCun (2016). Slides on Deep Learning Online (https://indico.cern.ch/event/510372/)
99. "Unsupervised Feature Learning and Deep Learning Tutorial" (http://ufldl.stanford.edu/tutorial/supervised/Convolution
alNeuralNetwork/).
100. Szegedy, Christian; Liu, Wei; Jia, Yangqing; Sermanet, Pierre; Reed, Scott; Anguelov, Dragomir; Erhan, Dumitru;
Vanhoucke, Vincent; Rabinovich, Andrew (2014). "Going Deeper with Convolutions". Computing Research
Repository: 1. arXiv:1409.4842 (https://arxiv.org/abs/1409.4842). doi:10.1109/CVPR.2015.7298594 (https://doi.org/1
0.1109%2FCVPR.2015.7298594). ISBN 978-1-4673-6964-0.
101. Ran, Lingyan; Zhang, Yanning; Zhang, Qilin; Yang, Tao (2017-06-12). "Convolutional Neural Network-Based Robot
Navigation Using Uncalibrated Spherical Images" (https://qilin-zhang.github.io/_pages/pdfs/sensors-17-01341.pdf)
(PDF). Sensors. MDPI AG. 17 (6): 1341. doi:10.3390/s17061341 (https://doi.org/10.3390%2Fs17061341).
102. Hinton, Geoffrey E.; Krizhevsky, Alex; Wang, Sida D. (2011), "Transforming Auto-Encoders" (https://link.springer.com/
chapter/10.1007%2F978-3-642-21735-7_6), Lecture Notes in Computer Science, Springer Berlin Heidelberg, pp. 44–
51, doi:10.1007/978-3-642-21735-7_6 (https://doi.org/10.1007%2F978-3-642-21735-7_6), ISBN 9783642217340,
retrieved 2018-09-27
103. Hochreiter, Sepp; Schmidhuber, Jürgen (1997-11-01). "Long Short-Term Memory" (http://www.mitpressjournals.org/do
i/10.1162/neco.1997.9.8.1735). Neural Computation. 9 (8): 1735–1780. doi:10.1162/neco.1997.9.8.1735 (https://doi.or
g/10.1162%2Fneco.1997.9.8.1735). ISSN 0899-7667 (https://www.worldcat.org/issn/0899-7667).
104. "Learning Precise Timing with LSTM Recurrent Networks (PDF Download Available)" (https://www.researchgate.net/p
ublication/220320057_Learning_Precise_Timing_with_LSTM_Recurrent_Networks). ResearchGate. pp. 115–143.
Retrieved 2017-06-13.
105. Bayer, Justin; Wierstra, Daan; Togelius, Julian; Schmidhuber, Jürgen (2009-09-14). "Evolving Memory Cell Structures
for Sequence Learning" (https://link.springer.com/chapter/10.1007/978-3-642-04277-5_76). Artificial Neural Networks
– ICANN 2009. Lecture Notes in Computer Science. Springer, Berlin, Heidelberg. 5769: 755–764. doi:10.1007/978-3-
642-04277-5_76 (https://doi.org/10.1007%2F978-3-642-04277-5_76). ISBN 978-3-642-04276-8.
106. Fernández, Santiago; Graves, Alex; Schmidhuber, Jürgen (2007). "Sequence labelling in structured domains with
hierarchical recurrent neural networks" (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.79.1887). In Proc.
20th Int. Joint Conf. on Artificial In℡ligence, Ijcai 2007: 774–779.
107. Graves, Alex; Fernández, Santiago; Gomez, Faustino (2006). "Connectionist temporal classification: Labelling
unsegmented sequence data with recurrent neural networks" (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.
1.1.75.6306). In Proceedings of the International Conference on Machine Learning, ICML 2006: 369–376.
108. Graves, Alex; Eck, Douglas; Beringer, Nicole; Schmidhuber, Jürgen (2003). "Biologically Plausible Speech
Recognition with LSTM Neural Nets" (Ftp://ftp.idsia.ch/pub/juergen/bioadit2004.pdf) (PDF). 1st Intl. Workshop on
Biologically Inspired Approaches to Advanced Information Technology, Bio-ADIT 2004, Lausanne, Switzerland.
pp. 175–184.
109. Fernández, Santiago; Graves, Alex; Schmidhuber, Jürgen (2007). "An Application of Recurrent Neural Networks to
Discriminative Keyword Spotting" (http://dl.acm.org/citation.cfm?id=1778066.1778092). Proceedings of the 17th
International Conference on Artificial Neural Networks. ICANN'07. Berlin, Heidelberg: Springer-Verlag: 220–229.
ISBN 3540746935.
110. Hannun, Awni; Case, Carl; Casper, Jared; Catanzaro, Bryan; Diamos, Greg; Elsen, Erich; Prenger, Ryan; Satheesh,
Sanjeev; Sengupta, Shubho (2014-12-17). "Deep Speech: Scaling up end-to-end speech recognition".
arXiv:1412.5567 (https://arxiv.org/abs/1412.5567) [cs.CL (https://arxiv.org/archive/cs.CL)].
111. Sak, Hasim; Senior, Andrew; Beaufays, Francoise (2014). "Long Short-Term Memory recurrent neural network
architectures for large scale acoustic modeling" (https://static.googleusercontent.com/media/research.google.com/e
n//pubs/archive/43905.pdf) (PDF).
112. Li, Xiangang; Wu, Xihong (2014-10-15). "Constructing Long Short-Term Memory based Deep Recurrent Neural
Networks for Large Vocabulary Speech Recognition". arXiv:1410.4281 (https://arxiv.org/abs/1410.4281) [cs.CL (http
s://arxiv.org/archive/cs.CL)].
113. Fan, Y.; Qian, Y.; Xie, F.; Soong, F. K. (2014). "TTS synthesis with bidirectional LSTM based Recurrent Neural
Networks" (https://www.researchgate.net/publication/287741874_TTS_synthesis_with_bidirectional_LSTM_based_R
ecurrent_Neural_Networks). ResearchGate. Retrieved 2017-06-13.
114. Zen, Heiga; Sak, Hasim (2015). "Unidirectional Long Short-Term Memory Recurrent Neural Network with Recurrent
Output Layer for Low-Latency Speech Synthesis" (https://static.googleusercontent.com/media/research.google.com/e
n//pubs/archive/43266.pdf) (PDF). Google.com. ICASSP. pp. 4470–4474.
115. Fan, Bo; Wang, Lijuan; Soong, Frank K.; Xie, Lei (2015). "Photo-Real Talking Head with Deep Bidirectional LSTM" (ht
tps://www.microsoft.com/en-us/research/wp-content/uploads/2015/04/icassp2015_fanbo_1009.pdf) (PDF).
Proceedings of ICASSP.
116. Sak, Haşim; Senior, Andrew; Rao, Kanishka; Beaufays, Françoise; Schalkwyk, Johan (September 2015). "Google
voice search: faster and more accurate" (http://googleresearch.blogspot.ch/2015/09/google-voice-search-faster-and-
more.html).
117. Gers, Felix A.; Schmidhuber, Jürgen (2001). "LSTM Recurrent Networks Learn Simple Context Free and Context
Sensitive Languages". IEEE Transactions on Neural Networks. 12 (6): 1333–1340. doi:10.1109/72.963769 (https://do
i.org/10.1109%2F72.963769).
118. A bot will complete this citation soon. Click here to jump the queue (http://tools.wmflabs.org/citations/process_page.php?page=Artifi
cial_neural_network) arXiv:1801.10111 (https://arxiv.org/abs/1801.10111).
119. Sutskever, L.; Vinyals, O.; Le, Q. (2014). "Sequence to Sequence Learning with Neural Networks" (https://papers.nip
s.cc/paper/5346-sequence-to-sequence-learning-with-neural-networks.pdf) (PDF). NIPS'14 Proceedings of the 27th
International Conference on Neural Information Processing Systems. 2: 3104–3112. arXiv:1409.3215 (https://arxiv.or
g/abs/1409.3215). Bibcode:2014arXiv1409.3215S (http://adsabs.harvard.edu/abs/2014arXiv1409.3215S).
120. Jozefowicz, Rafal; Vinyals, Oriol; Schuster, Mike; Shazeer, Noam; Wu, Yonghui (2016-02-07). "Exploring the Limits of
Language Modeling". arXiv:1602.02410 (https://arxiv.org/abs/1602.02410) [cs.CL (https://arxiv.org/archive/cs.CL)].
121. Gillick, Dan; Brunk, Cliff; Vinyals, Oriol; Subramanya, Amarnag (2015-11-30). "Multilingual Language Processing
From Bytes". arXiv:1512.00103 (https://arxiv.org/abs/1512.00103) [cs.CL (https://arxiv.org/archive/cs.CL)].
122. Vinyals, Oriol; Toshev, Alexander; Bengio, Samy; Erhan, Dumitru (2014-11-17). "Show and Tell: A Neural Image
Caption Generator". arXiv:1411.4555 (https://arxiv.org/abs/1411.4555) [cs.CV (https://arxiv.org/archive/cs.CV)].
123. Gallicchio, Claudio; Micheli, Alessio; Pedrelli, Luca (2017). "Deep reservoir computing: A critical experimental
analysis" (http://www.sciencedirect.com/science/article/pii/S0925231217307567). Neurocomputing. 268: 87–99.
doi:10.1016/j.neucom.2016.12.089 (https://doi.org/10.1016%2Fj.neucom.2016.12.089).
124. Gallicchio, Claudio; Micheli, Alessio (2017). "Echo State Property of Deep Reservoir Computing Networks". Cognitive
Computation. 9 (3): 337–350. doi:10.1007/s12559-017-9461-9 (https://doi.org/10.1007%2Fs12559-017-9461-9).
125. Hinton, G.E. (2009). "Deep belief networks". Scholarpedia. 4 (5): 5947. Bibcode:2009SchpJ...4.5947H (http://adsabs.
harvard.edu/abs/2009SchpJ...4.5947H). doi:10.4249/scholarpedia.5947 (https://doi.org/10.4249%2Fscholarpedia.594
7).
126. Larochelle, Hugo; Erhan, Dumitru; Courville, Aaron; Bergstra, James; Bengio, Yoshua (2007). "An Empirical
Evaluation of Deep Architectures on Problems with Many Factors of Variation" (http://doi.acm.org/10.1145/1273496.1
273556). Proceedings of the 24th International Conference on Machine Learning. ICML '07. New York, NY, USA:
ACM: 473–480. doi:10.1145/1273496.1273556 (https://doi.org/10.1145%2F1273496.1273556).
ISBN 9781595937933.
127. Graupe, Daniel (2013). Principles of Artificial Neural Networks (https://books.google.com/books?id=W6W6CgAAQBA
J&pg=PP1). World Scientific. pp. 1–. ISBN 978-981-4522-74-8.
128. A US 5920852 A (https://worldwide.espacenet.com/textdoc?DB=EPODOC&IDX=US5920852) D. Graupe," Large
memory storage and retrieval (LAMSTAR) network, April 1996
129. D. Graupe, "Principles of Artificial Neural Networks.3rd Edition", World Scientific Publishers, 2013, pp. 203–274.
130. Nigam, Vivek Prakash; Graupe, Daniel (2004-01-01). "A neural-network-based detection of epilepsy". Neurological
Research. 26 (1): 55–60. doi:10.1179/016164104773026534 (https://doi.org/10.1179%2F016164104773026534).
ISSN 0161-6412 (https://www.worldcat.org/issn/0161-6412). PMID 14977058 (https://www.ncbi.nlm.nih.gov/pubmed/1
4977058).
131. Waxman, Jonathan A.; Graupe, Daniel; Carley, David W. (2010-04-01). "Automated Prediction of Apnea and
Hypopnea, Using a LAMSTAR Artificial Neural Network" (http://www.atsjournals.org/doi/abs/10.1164/rccm.200907-11
46OC). American Journal of Respiratory and Critical Care Medicine. 181 (7): 727–733. doi:10.1164/rccm.200907-
1146oc (https://doi.org/10.1164%2Frccm.200907-1146oc). ISSN 1073-449X (https://www.worldcat.org/issn/1073-449
X).
132. Graupe, D.; Graupe, M. H.; Zhong, Y.; Jackson, R. K. (2008). "Blind adaptive filtering for non-invasive extraction of the
fetal electrocardiogram and its non-stationarities". Proc. Inst. Mech. Eng. H. 222 (8): 1221–1234.
doi:10.1243/09544119jeim417 (https://doi.org/10.1243%2F09544119jeim417).
133. Graupe 2013, pp. 240–253
134. Graupe, D.; Abon, J. (2002). "A Neural Network for Blind Adaptive Filtering of Unknown Noise from Speech" (https://w
ww.tib.eu/en/search/id/BLCP:CN019373941/Blind-Adaptive-Filtering-of-Speech-from-Noise-of/). Intelligent
Engineering Systems Through Artificial Neural Networks. Technische Informationsbibliothek (TIB). 12: 683–688.
Retrieved 2017-06-14.
135. D. Graupe, "Principles of Artificial Neural Networks.3rd Edition", World Scientific Publishers", 2013, pp. 253–274.
136. Girado, J. I.; Sandin, D. J.; DeFanti, T. A. (2003). "Real-time camera-based face detection using a modified LAMSTAR
neural network system". Proc. SPIE 5015, Applications of Artificial Neural Networks in Image Processing VIII.
Applications of Artificial Neural Networks in Image Processing VIII. 5015: 36. Bibcode:2003SPIE.5015...36G (http://ad
sabs.harvard.edu/abs/2003SPIE.5015...36G). doi:10.1117/12.477405 (https://doi.org/10.1117%2F12.477405).
137. Venkatachalam, V; Selvan, S. (2007). "Intrusion Detection using an Improved Competitive Learning Lamstar
Network". International Journal of Computer Science and Network Security. 7 (2): 255–263.
138. Graupe, D.; Smollack, M. (2007). "Control of unstable nonlinear and nonstationary systems using LAMSTAR neural
networks" (https://www.researchgate.net/publication/262316982_Control_of_unstable_nonlinear_and_nonstationary_
systems_using_LAMSTAR_neural_networks). ResearchGate. Proceedings of 10th IASTED on Intelligent Control,
Sect.592,. pp. 141–144. Retrieved 2017-06-14.
139. Graupe, Daniel (7 July 2016). Deep Learning Neural Networks: Design and Case Studies (https://books.google.com/b
ooks?id=e5hIDQAAQBAJ&pg=PA57). World Scientific Publishing Co Inc. pp. 57–110. ISBN 978-981-314-647-1.
140. Graupe, D.; Kordylewski, H. (August 1996). "Network based on SOM (Self-Organizing-Map) modules combined with
statistical decision tools" (http://ieeexplore.ieee.org/document/594203/). Proceedings of the 39th Midwest Symposium
on Circuits and Systems. 1: 471–474 vol.1. doi:10.1109/mwscas.1996.594203 (https://doi.org/10.1109%2Fmwscas.19
96.594203). ISBN 0-7803-3636-4.
141. Graupe, D.; Kordylewski, H. (1998-03-01). "A Large Memory Storage and Retrieval Neural Network for Adaptive
Retrieval and Diagnosis" (http://www.worldscientific.com/doi/abs/10.1142/S0218194098000091). International Journal
of Software Engineering and Knowledge Engineering. 08 (1): 115–138. doi:10.1142/s0218194098000091 (https://doi.
org/10.1142%2Fs0218194098000091). ISSN 0218-1940 (https://www.worldcat.org/issn/0218-1940).
142. Kordylewski, H.; Graupe, D; Liu, K. (2001). "A novel large-memory neural network as an aid in medical diagnosis
applications". IEEE Transactions on Information Technology in Biomedicine. 5 (3): 202–209.
doi:10.1109/4233.945291 (https://doi.org/10.1109%2F4233.945291).
143. Schneider, N.C.; Graupe (2008). "A modified LAMSTAR neural network and its applications". International journal of
neural systems. 18 (4): 331–337. doi:10.1142/s0129065708001634
(https://doi.org/10.1142%2Fs0129065708001634).
144. Graupe 2013, p. 217
145. Vincent, Pascal; Larochelle, Hugo; Lajoie, Isabelle; Bengio, Yoshua; Manzagol, Pierre-Antoine (2010). "Stacked
Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion" (htt
p://dl.acm.org/citation.cfm?id=1953039). The Journal of Machine Learning Research. 11: 3371–3408.
146. Ballard, Dana H. (1987). "Modular learning in neural networks" (http://www.aaai.org/Papers/AAAI/1987/AAAI87-050.p
df) (PDF). Proceedings of AAAI. pp. 279–284.
147. Deng, Li; Yu, Dong; Platt, John (2012). "Scalable stacking and learning for building deep architectures" (http://researc
h-srv.microsoft.com/pubs/157586/DSN-ICASSP2012.pdf) (PDF). 2012 IEEE International Conference on Acoustics,
Speech and Signal Processing (ICASSP): 2133–2136.
148. Deng, Li; Yu, Dong (2011). "Deep Convex Net: A Scalable Architecture for Speech Pattern Classification" (http://www.
truebluenegotiations.com/files/deepconvexnetwork-interspeech2011-pub.pdf) (PDF). Proceedings of the Interspeech:
2285–2288.
149. David, Wolpert (1992). "Stacked generalization". Neural Networks. 5 (2): 241–259. doi:10.1016/S0893-
6080(05)80023-1 (https://doi.org/10.1016%2FS0893-6080%2805%2980023-1).
150. Bengio, Y. (2009-11-15). "Learning Deep Architectures for AI" (http://www.nowpublishers.com/article/Details/MAL-
006). Foundations and Trends® in Machine Learning. 2 (1): 1–127. doi:10.1561/2200000006 (https://doi.org/10.156
1%2F2200000006). ISSN 1935-8237 (https://www.worldcat.org/issn/1935-8237).
151. Hutchinson, Brian; Deng, Li; Yu, Dong (2012). "Tensor deep stacking networks". IEEE Transactions on Pattern
Analysis and Machine Intelligence. 1–15 (8): 1944–1957. doi:10.1109/tpami.2012.268 (https://doi.org/10.1109%2Ftpa
mi.2012.268).
152. Hinton, Geoffrey; Salakhutdinov, Ruslan (2006). "Reducing the Dimensionality of Data with Neural Networks".
Science. 313 (5786): 504–507. Bibcode:2006Sci...313..504H (http://adsabs.harvard.edu/abs/2006Sci...313..504H).
doi:10.1126/science.1127647 (https://doi.org/10.1126%2Fscience.1127647). PMID 16873662 (https://www.ncbi.nlm.ni
h.gov/pubmed/16873662).
153. Dahl, G.; Yu, D.; Deng, L.; Acero, A. (2012). "Context-Dependent Pre-Trained Deep Neural Networks for Large-
Vocabulary Speech Recognition". IEEE Transactions on Audio, Speech, and Language Processing. 20 (1): 30–42.
doi:10.1109/tasl.2011.2134090 (https://doi.org/10.1109%2Ftasl.2011.2134090).
154. Mohamed, Abdel-rahman; Dahl, George; Hinton, Geoffrey (2012). "Acoustic Modeling Using Deep Belief Networks".
IEEE Transactions on Audio, Speech, and Language Processing. 20 (1): 14–22. doi:10.1109/tasl.2011.2109382 (http
s://doi.org/10.1109%2Ftasl.2011.2109382).
155. Courville, Aaron; Bergstra, James; Bengio, Yoshua (2011). "A Spike and Slab Restricted Boltzmann Machine" (http://
machinelearning.wustl.edu/mlpapers/paper_files/AISTATS2011_CourvilleBB11.pdf) (PDF). JMLR: Workshop and
Conference Proceeding. 15: 233–241.
156. Courville, Aaron; Bergstra, James; Bengio, Yoshua (2011). "Unsupervised Models of Images by Spike-and-Slab
RBMs". Proceedings of the 28th International Conference on Machine Learning (http://machinelearning.wustl.edu/mlp
apers/paper_files/ICML2011Courville_591.pdf) (PDF). 10. pp. 1–8.
157. Mitchell, T; Beauchamp, J (1988). "Bayesian Variable Selection in Linear Regression". Journal of the American
Statistical Association. 83 (404): 1023–1032. doi:10.1080/01621459.1988.10478694 (https://doi.org/10.1080%2F016
21459.1988.10478694).
158. Hinton, Geoffrey; Salakhutdinov, Ruslan (2009). "Efficient Learning of Deep Boltzmann Machines" (http://machinelear
ning.wustl.edu/mlpapers/paper_files/AISTATS09_SalakhutdinovH.pdf) (PDF). 3: 448–455.
159. Larochelle, Hugo; Bengio, Yoshua; Louradour, Jerdme; Lamblin, Pascal (2009). "Exploring Strategies for Training
Deep Neural Networks" (http://dl.acm.org/citation.cfm?id=1577070). The Journal of Machine Learning Research. 10:
1–40.
160. Coates, Adam; Carpenter, Blake (2011). "Text Detection and Character Recognition in Scene Images with
Unsupervised Feature Learning" (http://www.iapr-tc11.org/archive/icdar2011/fileup/PDF/4520a440.pdf) (PDF): 440–
445.
161. Lee, Honglak; Grosse, Roger (2009). "Convolutional deep belief networks for scalable unsupervised learning of
hierarchical representations" (http://portal.acm.org/citation.cfm?doid=1553374.1553453). Proceedings of the 26th
Annual International Conference on Machine Learning: 1–8.
162. Lin, Yuanqing; Zhang, Tong (2010). "Deep Coding Network" (http://machinelearning.wustl.edu/mlpapers/paper_files/N
IPS2010_1077.pdf) (PDF). Advances in Neural . . .: 1–9.
163. Ranzato, Marc Aurelio; Boureau, Y-Lan (2007). "Sparse Feature Learning for Deep Belief Networks" (http://machinele
arning.wustl.edu/mlpapers/paper_files/NIPS2007_1118.pdf) (PDF). Advances in Neural Information Processing
Systems. 23: 1–8.
164. Socher, Richard; Lin, Clif (2011). "Parsing Natural Scenes and Natural Language with Recursive Neural Networks" (ht
tp://machinelearning.wustl.edu/mlpapers/paper_files/ICML2011Socher_125.pdf) (PDF). Proceedings of the 26th
International Conference on Machine Learning.
165. Taylor, Graham; Hinton, Geoffrey (2006). "Modeling Human Motion Using Binary Latent Variables" (http://machinelear
ning.wustl.edu/mlpapers/paper_files/NIPS2006_693.pdf) (PDF). Advances in Neural Information Processing Systems.
166. Vincent, Pascal; Larochelle, Hugo (2008). "Extracting and composing robust features with denoising autoencoders" (h
ttp://portal.acm.org/citation.cfm?doid=1390156.1390294). Proceedings of the 25th international conference on
Machine learning – ICML '08: 1096–1103.
167. Kemp, Charles; Perfors, Amy; Tenenbaum, Joshua (2007). "Learning overhypotheses with hierarchical Bayesian
models". Developmental Science. 10 (3): 307–21. doi:10.1111/j.1467-7687.2007.00585.x (https://doi.org/10.1111%2Fj.
1467-7687.2007.00585.x). PMID 17444972 (https://www.ncbi.nlm.nih.gov/pubmed/17444972).
168. Xu, Fei; Tenenbaum, Joshua (2007). "Word learning as Bayesian inference". Psychol. Rev. 114 (2): 245–72.
doi:10.1037/0033-295X.114.2.245 (https://doi.org/10.1037%2F0033-295X.114.2.245). PMID 17500627 (https://www.n
cbi.nlm.nih.gov/pubmed/17500627).
169. Chen, Bo; Polatkan, Gungor (2011). "The Hierarchical Beta Process for Convolutional Factor Analysis and Deep
Learning" (http://machinelearning.wustl.edu/mlpapers/paper_files/ICML2011Chen_251.pdf) (PDF). Machine Learning
...
170. Fei-Fei, Li; Fergus, Rob (2006). "One-shot learning of object categories". IEEE Transactions on Pattern Analysis and
Machine Intelligence. 28 (4): 594–611. doi:10.1109/TPAMI.2006.79 (https://doi.org/10.1109%2FTPAMI.2006.79).
PMID 16566508 (https://www.ncbi.nlm.nih.gov/pubmed/16566508).
171. Rodriguez, Abel; Dunson, David (2008). "The Nested Dirichlet Process" (http://amstat.tandfonline.com/doi/full/10.119
8/016214508000000553). Journal of the American Statistical Association. 103 (483): 1131–1154.
172. Ruslan, Salakhutdinov; Joshua, Tenenbaum (2012). "Learning with Hierarchical-Deep Models". IEEE Transactions on
Pattern Analysis and Machine Intelligence. 35 (8): 1958–71. doi:10.1109/TPAMI.2012.269 (https://doi.org/10.1109%2
FTPAMI.2012.269). PMID 23787346 (https://www.ncbi.nlm.nih.gov/pubmed/23787346).
173. Chalasani, Rakesh; Principe, Jose (2013). "Deep Predictive Coding Networks". arXiv:1301.3541 (https://arxiv.org/abs/
1301.3541) [cs.LG (https://arxiv.org/archive/cs.LG)].
174. Hinton, Geoffrey E. (1984). "Distributed representations" (http://repository.cmu.edu/cgi/viewcontent.cgi?article=2841&
context=compsci).
175. S. Das, C.L. Giles, G.Z. Sun, "Learning Context Free Grammars: Limitations of a Recurrent Neural Network with an
External Stack Memory," Proc. 14th Annual Conf. of the Cog. Sci. Soc., p. 79, 1992.
176. Mozer, M. C.; Das, S. (1993). "A connectionist symbol manipulator that discovers the structure of context-free
languages" (https://papers.nips.cc/paper/626-a-connectionist-symbol-manipulator-that-discovers-the-structure-of-cont
ext-free-languages). NIPS 5. pp. 863–870.
177. Schmidhuber, J. (1992). "Learning to control fast-weight memories: An alternative to recurrent nets". Neural
Computation. 4 (1): 131–139. doi:10.1162/neco.1992.4.1.131 (https://doi.org/10.1162%2Fneco.1992.4.1.131).
178. Gers, F.; Schraudolph, N.; Schmidhuber, J. (2002). "Learning precise timing with LSTM recurrent networks" (http://jml
r.org/papers/volume3/gers02a/gers02a.pdf) (PDF). JMLR. 3: 115–143.
179. Jürgen Schmidhuber (1993). "An introspective network that can learn to run its own weight change algorithm" (ftp://ft
p.idsia.ch/pub/juergen/iee93self.ps.gz). In Proc. of the Intl. Conf. on Artificial Neural Networks, Brighton. IEE.
pp. 191–195.
180. Hochreiter, Sepp; Younger, A. Steven; Conwell, Peter R. (2001). "Learning to Learn Using Gradient Descent" (http://ci
teseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.5.323). ICANN. 2130: 87–94.
181. https://arxiv.org/pdf/1506.02516.pdf
183. Burgess, Matt. "DeepMind's AI learned to ride the London Underground using human-like reason and memory" (http
s://www.wired.co.uk/article/deepmind-ai-tube-london-underground). WIRED UK. Retrieved 2016-10-19.
184. "DeepMind AI 'Learns' to Navigate London Tube" (https://www.pcmag.com/news/348701/deepmind-ai-learns-to-navig
ate-london-tube). PCMAG. Retrieved 2016-10-19.
185. Mannes, John. "DeepMind's differentiable neural computer helps you navigate the subway with its memory" (https://te
chcrunch.com/2016/10/13/__trashed-2/). TechCrunch. Retrieved 2016-10-19.
186. Graves, Alex; Wayne, Greg; Reynolds, Malcolm; Harley, Tim; Danihelka, Ivo; Grabska-Barwińska, Agnieszka;
Colmenarejo, Sergio Gómez; Grefenstette, Edward; Ramalho, Tiago (2016-10-12). "Hybrid computing using a neural
network with dynamic external memory" (http://www.nature.com/nature/journal/vaop/ncurrent/full/nature20101.html).
Nature. 538 (7626): 471–476. Bibcode:2016Natur.538..471G (http://adsabs.harvard.edu/abs/2016Natur.538..471G).
doi:10.1038/nature20101 (https://doi.org/10.1038%2Fnature20101). ISSN 1476-4687 (https://www.worldcat.org/issn/1
476-4687). PMID 27732574 (https://www.ncbi.nlm.nih.gov/pubmed/27732574).
187. "Differentiable neural computers | DeepMind" (https://deepmind.com/blog/differentiable-neural-computers/).
DeepMind. Retrieved 2016-10-19.
188. Atkeson, Christopher G.; Schaal, Stefan (1995). "Memory-based neural networks for robot learning".
Neurocomputing. 9 (3): 243–269. doi:10.1016/0925-2312(95)00033-6 (https://doi.org/10.1016%2F0925-2312%289
5%2900033-6).
189. Salakhutdinov, Ruslan, and Geoffrey Hinton. "Semantic hashing." (http://www.utstat.toronto.edu/~rsalakhu/papers/sda
rticle.pdf) International Journal of Approximate Reasoning 50.7 (2009): 969–978.
190. Le, Quoc V.; Mikolov, Tomas (2014). "Distributed representations of sentences and documents". arXiv:1405.4053 (htt
ps://arxiv.org/abs/1405.4053) [cs.CL (https://arxiv.org/archive/cs.CL)].
194. "AI device identifies objects at the speed of light: The 3D-printed artificial neural network can be used in medicine,
robotics and security" (https://www.sciencedaily.com/releases/2018/08/180802130750.htm). ScienceDaily. Retrieved
2018-08-08.
197. Kalchbrenner, N.; Blunsom, P. (2013). "Recurrent continuous translation models" (http://www.aclweb.org/anthology/D
13-1176). EMNLP'2013.
198. Sutskever, I.; Vinyals, O.; Le, Q. V. (2014). "Sequence to sequence learning with neural networks" (https://papers.nip
s.cc/paper/5346-sequence-to-sequence-learning-with-neural-networks.pdf) (PDF). NIPS'2014.
201. Scholkopf, B; Smola, Alexander (1998). "Nonlinear component analysis as a kernel eigenvalue problem". Neural
computation. (44) (5): 1299–1319. CiteSeerX 10.1.1.53.8911 (https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.
1.1.53.8911). doi:10.1162/089976698300017467 (https://doi.org/10.1162%2F089976698300017467).
202. Cho, Youngmin (2012). "Kernel Methods for Deep Learning" (http://cseweb.ucsd.edu/~yoc002/paper/thesis_youngmin
cho.pdf) (PDF): 1–9.
203. Deng, Li; Tur, Gokhan; He, Xiaodong; Hakkani-Tür, Dilek (2012-12-01). "Use of Kernel Deep Convex Networks and
End-To-End Learning for Spoken Language Understanding" (https://www.microsoft.com/en-us/research/publication/us
e-of-kernel-deep-convex-networks-and-end-to-end-learning-for-spoken-language-understanding/). Microsoft
Research.
204. Zoph, Barret; Le, Quoc V. (2016-11-04). "Neural Architecture Search with Reinforcement Learning". arXiv:1611.01578
(https://arxiv.org/abs/1611.01578) [cs.LG (https://arxiv.org/archive/cs.LG)].
205. Zissis, Dimitrios (October 2015). "A cloud based architecture capable of perceiving and predicting multiple vessel
behaviour" (http://www.sciencedirect.com/science/article/pii/S1568494615004329). Applied Soft Computing. 35: 652–
661. doi:10.1016/j.asoc.2015.07.002 (https://doi.org/10.1016%2Fj.asoc.2015.07.002).
206. Roman M. Balabin; Ekaterina I. Lomakina (2009). "Neural network approach to quantum-chemistry data: Accurate
prediction of density functional theory energies". J. Chem. Phys. 131 (7): 074104. Bibcode:2009JChPh.131g4104B (h
ttp://adsabs.harvard.edu/abs/2009JChPh.131g4104B). doi:10.1063/1.3206326 (https://doi.org/10.1063%2F1.320632
207. Sengupta, Nandini; Sahidullah, Md; Saha, Goutam (August 2016). "Lung sound classification using cepstral-based
statistical features" (http://www.sciencedirect.com/science/article/pii/S0010482516301263). Computers in Biology and
Medicine. 75 (1): 118–129. doi:10.1016/j.compbiomed.2016.05.013 (https://doi.org/10.1016%2Fj.compbiomed.2016.0
5.013).
208. French, Jordan (2016). "The time traveller's CAPM" (http://www.tandfonline.com/doi/abs/10.1080/10293523.2016.125
5469). Investment Analysts Journal. 46 (2): 81–96. doi:10.1080/10293523.2016.1255469 (https://doi.org/10.1080%2F
10293523.2016.1255469).
209. Schechner, Sam (2017-06-15). "Facebook Boosts A.I. to Block Terrorist Propaganda" (https://www.wsj.com/articles/fa
cebook-boosts-a-i-to-block-terrorist-propaganda-1497546000). Wall Street Journal. ISSN 0099-9660 (https://www.wor
ldcat.org/issn/0099-9660). Retrieved 2017-06-16.
210. Ganesan, N. "Application of Neural Networks in Diagnosing Cancer Disease Using Demographic Data" (http://www.ijc
aonline.org/journal/number26/pxc387783.pdf) (PDF). International Journal of Computer Applications.
211. Bottaci, Leonardo. "Artificial Neural Networks Applied to Outcome Prediction for Colorectal Cancer Patients in
Separate Institutions" (http://www.lcc.uma.es/~jja/recidiva/042.pdf) (PDF). The Lancet.
212. Alizadeh, Elaheh; Lyons, Samanthe M; Castle, Jordan M; Prasad, Ashok (2016). "Measuring systematic changes in
invasive cancer cell shape using Zernike moments" (http://pubs.rsc.org/en/Content/ArticleLanding/2016/IB/C6IB0010
0A#!divAbstract). Integrative Biology. 8 (11): 1183–1193. doi:10.1039/C6IB00100A (https://doi.org/10.1039%2FC6IB0
0100A). PMID 27735002 (https://www.ncbi.nlm.nih.gov/pubmed/27735002).
213. Lyons, Samanthe (2016). "Changes in cell shape are correlated with metastatic potential in murine" (http://bio.biologis
ts.org/content/5/3/289). Biology Open. 5 (3): 289–299. doi:10.1242/bio.013409 (https://doi.org/10.1242%2Fbio.01340
9).
214. Nabian, Mohammad Amin; Meidani, Hadi (2017-08-28). "Deep Learning for Accelerated Reliability Analysis of
Infrastructure Networks". arXiv:1708.08551 (https://arxiv.org/abs/1708.08551) [cs.CE
(https://arxiv.org/archive/cs.CE)].
215. Nabian, Mohammad Amin; Meidani, Hadi (2018). "Accelerating Stochastic Assessment of Post-Earthquake
Transportation Network Connectivity via Machine-Learning-Based Surrogates" (https://trid.trb.org/view/1496617).
Transportation Research Board 97th Annual Meeting.
216. null null (2000-04-01). "Artificial Neural Networks in Hydrology. I: Preliminary Concepts" (http://ascelibrary.org/doi/abs/
10.1061/(ASCE)1084-0699(2000)5:2(115)). Journal of Hydrologic Engineering. 5 (2): 115–123.
doi:10.1061/(ASCE)1084-0699(2000)5:2(115) (https://doi.org/10.1061%2F%28ASCE%291084-0699%282000%295%
3A2%28115%29).
217. null null (2000-04-01). "Artificial Neural Networks in Hydrology. II: Hydrologic Applications" (http://ascelibrary.org/doi/a
bs/10.1061/(ASCE)1084-0699(2000)5:2(124)). Journal of Hydrologic Engineering. 5 (2): 124–137.
doi:10.1061/(ASCE)1084-0699(2000)5:2(124) (https://doi.org/10.1061%2F%28ASCE%291084-0699%282000%295%
3A2%28124%29).
218. Peres, D. J.; Iuppa, C.; Cavallaro, L.; Cancelliere, A.; Foti, E. (2015-10-01). "Significant wave height record extension
by neural networks and reanalysis wind data" (http://www.sciencedirect.com/science/article/pii/S1463500315001432).
Ocean Modelling. 94: 128–140. Bibcode:2015OcMod..94..128P (http://adsabs.harvard.edu/abs/2015OcMod..94..128
P). doi:10.1016/j.ocemod.2015.08.002 (https://doi.org/10.1016%2Fj.ocemod.2015.08.002).
219. Dwarakish, G. S.; Rakshith, Shetty; Natesan, Usha (2013). "Review on Applications of Neural Network in Coastal
Engineering" (http://www.ciitresearch.org/dl/index.php/aiml/article/view/AIML072013007). Artificial Intelligent Systems
and Machine Learning. 5 (7): 324–331.
220. Ermini, Leonardo; Catani, Filippo; Casagli, Nicola (2005-03-01). "Artificial Neural Networks applied to landslide
susceptibility assessment" (http://www.sciencedirect.com/science/article/pii/S0169555X04002272). Geomorphology.
Geomorphological hazard and human impact in mountain environments. 66 (1): 327–343.
Bibcode:2005Geomo..66..327E (http://adsabs.harvard.edu/abs/2005Geomo..66..327E).
doi:10.1016/j.geomorph.2004.09.025 (https://doi.org/10.1016%2Fj.geomorph.2004.09.025).
221. Forrest MD (April 2015). "Simulation of alcohol action upon a detailed Purkinje neuron model and a simpler surrogate
model that runs >400 times faster" (http://www.biomedcentral.com/1471-2202/16/27). BMC Neuroscience. 16 (27).
doi:10.1186/s12868-015-0162-6 (https://doi.org/10.1186%2Fs12868-015-0162-6).
222. Siegelmann, H.T.; Sontag, E.D. (1991). "Turing computability with neural nets" (http://www.math.rutgers.edu/~sontag/
FTPDIR/aml-turing.pdf) (PDF). Appl. Math. Lett. 4 (6): 77–80. doi:10.1016/0893-9659(91)90080-F (https://doi.org/10.1
016%2F0893-9659%2891%2990080-F).
223. Balcázar, José (Jul 1997). "Computational Power of Neural Networks: A Kolmogorov Complexity Characterization" (ht
tp://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=605580&url=http%3A%2F%2Fieeexplore.ieee.org%2Fxpls%2Fa
bs_all.jsp%3Farnumber%3D605580). Information Theory, IEEE Transactions on. 43 (4): 1175–1183.
CiteSeerX 10.1.1.411.7782 (https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.411.7782).
doi:10.1109/18.605580 (https://doi.org/10.1109%2F18.605580). Retrieved 3 November 2014.
224. Crick, Francis (1989). "The recent excitement about neural networks" (http://europepmc.org/abstract/med/2911347).
Nature. 337 (6203): 129–132. Bibcode:1989Natur.337..129C (http://adsabs.harvard.edu/abs/1989Natur.337..129C).
doi:10.1038/337129a0 (https://doi.org/10.1038%2F337129a0). PMID 2911347 (https://www.ncbi.nlm.nih.gov/pubmed/
2911347).
225. Adrian, Edward D. (1926). "The impulses produced by sensory nerve endings" (http://onlinelibrary.wiley.com/doi/10.11
13/jphysiol.1926.sp002273/full). The Journal of Physiology. 61 (1): 49–72. doi:10.1113/jphysiol.1926.sp002273 (http
s://doi.org/10.1113%2Fjphysiol.1926.sp002273). PMC 1514809 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1514
226. Dewdney, A. K. (1 April 1997). Yes, we have no neutrons: an eye-opening tour through the twists and turns of bad
science (https://books.google.com/books?id=KcHaAAAAMAAJ&pg=PA82). Wiley. p. 82. ISBN 978-0-471-10806-1.
227. D. J. Felleman and D. C. Van Essen, "Distributed hierarchical processing in the primate cerebral cortex (http://cercor.o
xfordjournals.org/content/1/1/1.1.full.pdf+html)," Cerebral Cortex, 1, pp. 1–47, 1991.
228. J. Weng, "Natural and Artificial Intelligence: Introduction to Computational Brain-Mind (https://www.amazon.com/Natur
al-Artificial-Intelligence-Introduction-Computational/dp/0985875720)," BMI Press, ISBN 978-0985875725, 2012.
229. Edwards, Chris (25 June 2015). "Growing pains for deep learning". Communications of the ACM. 58 (7): 14–16.
230. Schmidhuber, Jürgen (2015). "Deep learning in neural networks: An overview". Neural Networks. 61: 85–117.
arXiv:1404.7828 (https://arxiv.org/abs/1404.7828). doi:10.1016/j.neunet.2014.09.003 (https://doi.org/10.1016%2Fj.ne
unet.2014.09.003). PMID 25462637 (https://www.ncbi.nlm.nih.gov/pubmed/25462637).
231. Cade Metz (May 18, 2016). "Google Built Its Very Own Chips to Power Its AI Bots" (https://www.wired.com/2016/05/g
oogle-tpu-custom-chips/). Wired.
232. NASA – Dryden Flight Research Center – News Room: News Releases: NASA NEURAL NETWORK PROJECT
PASSES MILESTONE (http://www.nasa.gov/centers/dryden/news/NewsReleases/2003/03-49.html). Nasa.gov.
Retrieved on 2013-11-20.
233. "Roger Bridgman's defence of neural networks" (https://web.archive.org/web/20120319163352/http://members.fortun
ecity.com/templarseries/popper.html). Archived from the original (http://members.fortunecity.com/templarseries/poppe
r.html) on 19 March 2012. Retrieved 12 July 2010.
234. "Scaling Learning Algorithms towards {AI} – LISA – Publications – Aigaion 2.0" (http://www.iro.umontreal.ca/~lisa/publ
ications2/index.php/publications/show/4).
235. Sun and Bookman (1990)
236. Tahmasebi; Hezarkhani (2012). "A hybrid neural networks-fuzzy logic-genetic algorithm for grade estimation" (http://w
ww.sciencedirect.com/science/article/pii/S0098300412000398). Computers & Geosciences. 42: 18–27.
Bibcode:2012CG.....42...18T (http://adsabs.harvard.edu/abs/2012CG.....42...18T). doi:10.1016/j.cageo.2012.02.004
(https://doi.org/10.1016%2Fj.cageo.2012.02.004).
Bibliography
Bhadeshia H. K. D. H. (1999). "Neural Networks in Materials Science" (http://www.msm.cam.ac.uk/phase-trans/abstra
cts/neural.review.pdf) (PDF). ISIJ International. 39 (10): 966–979. doi:10.2355/isijinternational.39.966 (https://doi.org/
10.2355%2Fisijinternational.39.966).
M., Bishop, Christopher (1995). Neural networks for pattern recognition (https://www.worldcat.org/oclc/33101074).
Clarendon Press. ISBN 0198538499. OCLC 33101074 (https://www.worldcat.org/oclc/33101074).
Borgelt,, Christian (2003). Neuro-Fuzzy-Systeme : von den Grundlagen künstlicher Neuronaler Netze zur Kopplung
mit Fuzzy-Systemen (https://www.worldcat.org/oclc/76538146). Vieweg. ISBN 9783528252656. OCLC 76538146 (htt
ps://www.worldcat.org/oclc/76538146).
Cybenko, G.V. (2006). "Approximation by Superpositions of a Sigmoidal function" (https://books.google.com/books?id
=4RtVAAAAMAAJ&pg=PA303). In van Schuppen, Jan H. Mathematics of Control, Signals, and Systems. Springer
International. pp. 303–314. PDF (https://web.archive.org/web/20110719183058/http://actcomm.dartmouth.edu/gvc/pa
pers/approx_by_superposition.pdf)
Dewdney, A. K. (1997). Yes, we have no neutrons : an eye-opening tour through the twists and turns of bad science
(https://www.worldcat.org/oclc/35558945). New York: Wiley. ISBN 9780471108061. OCLC 35558945 (https://www.wor
ldcat.org/oclc/35558945).
Duda, Richard O.; Hart, Peter Elliot; Stork, David G. (2001). Pattern classification (https://www.worldcat.org/oclc/4134
7061) (2 ed.). Wiley. ISBN 0471056693. OCLC 41347061 (https://www.worldcat.org/oclc/41347061).
Egmont-Petersen, M.; de Ridder, D.; Handels, H. (2002). "Image processing with neural networks – a review". Pattern
Recognition. 35 (10): 2279–2301. doi:10.1016/S0031-3203(01)00178-9 (https://doi.org/10.1016%2FS0031-3203%28
01%2900178-9).
Fahlman, S.; Lebiere, C (1991). "The Cascade-Correlation Learning Architecture" (http://www.cs.iastate.edu/~honava
r/fahlman.pdf) (PDF).
created for National Science Foundation, Contract Number EET-8716324, and Defense Advanced Research
Projects Agency (DOD), ARPA Order No. 4976 under Contract F33615-87-C-1499.
Gurney, Kevin (1997). An introduction to neural networks (https://www.worldcat.org/oclc/37875698). UCL Press.
ISBN 1857286731. OCLC 37875698 (https://www.worldcat.org/oclc/37875698).
Haykin, Simon S. (1999). Neural networks : a comprehensive foundation (https://www.worldcat.org/oclc/38908586).
Prentice Hall. ISBN 0132733501. OCLC 38908586 (https://www.worldcat.org/oclc/38908586).
Hertz, J.; Palmer, Richard G.; Krogh, Anders S. (1991). Introduction to the theory of neural computation (https://www.
worldcat.org/oclc/21522159). Addison-Wesley. ISBN 0201515601. OCLC 21522159 (https://www.worldcat.org/oclc/21
522159).
Information theory, inference, and learning algorithms. Cambridge University Press. ISBN 9780521642989.
OCLC 52377690 (https://www.worldcat.org/oclc/52377690).
Kruse, Rudolf,; Borgelt, Christian; Klawonn, F.; Moewes, Christian; Steinbrecher, Matthias; Held,, Pascal (2013).
Computational intelligence : a methodological introduction (https://www.worldcat.org/oclc/837524179). Springer.
ISBN 9781447150121. OCLC 837524179 (https://www.worldcat.org/oclc/837524179).
Lawrence, Jeanette (1994). Introduction to neural networks : design, theory and applications (https://www.worldcat.or
g/oclc/32179420). California Scientific Software. ISBN 1883157005. OCLC 32179420 (https://www.worldcat.org/oclc/3
2179420).
MacKay, David, J.C. (2003). Information Theory, Inference, and Learning Algorithms (http://www.inference.phy.cam.a
c.uk/itprnn/book.pdf) (PDF). Cambridge University Press. ISBN 9780521642989.
Masters,, Timothy (1994). Signal and image processing with neural networks : a C++ sourcebook (https://www.worldc
at.org/oclc/29877717). J. Wiley. ISBN 0471049638. OCLC 29877717 (https://www.worldcat.org/oclc/29877717).
Ripley, Brian D. (2007). Pattern Recognition and Neural Networks (https://books.google.com/books?id=m12UR8QmL
qoC). Cambridge University Press. ISBN 978-0-521-71770-0.
Siegelmann, H.T.; Sontag, Eduardo D. (1994). "Analog computation via neural networks" (https://pdfs.semanticschola
r.org/861e/de32115d157e1568622b153e7ed3dca28467.pdf) (PDF). Theoretical Computer Science. 131 (2): 331–360.
doi:10.1016/0304-3975(94)90178-3 (https://doi.org/10.1016%2F0304-3975%2894%2990178-3).
Smith, Murray (1993). Neural networks for statistical modeling (https://www.worldcat.org/oclc/27145760). Van
Nostrand Reinhold. ISBN 0442013108. OCLC 27145760 (https://www.worldcat.org/oclc/27145760).
Wasserman, Philip D. (1993). Advanced methods in neural computing (https://www.worldcat.org/oclc/27429729). Van
Nostrand Reinhold. ISBN 0442004613. OCLC 27429729 (https://www.worldcat.org/oclc/27429729).
External links
A brief introduction to Neural Networks (http://www.dkriesel.com/en/science/neural_networks) (PDF), illustrated 250p
textbook covering the common kinds of neural networks (CC license).
An Introduction to Deep Neural Networks (https://web.archive.org/web/20150726190848/http://deeplearning4j.org/neu
ralnet-overview.html).
A Tutorial of Neural Network in Excel (http://people.revoledu.com/kardi/tutorial/NeuralNetwork/index.html).
MIT course on Neural Networks (https://www.youtube.com/watch?v=q0pm3BrIUFo) on YouTube
A Concise Introduction to Machine Learning with Artificial Neural Networks (https://www.academia.edu/25708860/A_
Concise_Introduction_to_Machine_Learning_with_Artificial_Neural_Networks)
Neural Networks for Machine Learning – a course by Geoffrey Hinton (https://www.coursera.org/course/neuralnets)
Deep Learning (http://www.deeplearningbook.org/)
Interactive visualization of neural work (https://nnplayground.com)
How Artificial Neural Network Types Can Change Business (https://biztechmagazine.com/article/2018/09/how-artificial
-neural-network-types-can-change-business-perfcon)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Artificial_neural_network&oldid=862148531"
This page was last edited on 2 October 2018, at 13:46 (UTC).
Text is available under the Creative Commons Attribution-ShareAlike License; additional terms may apply. By using this
site, you agree to the Terms of Use and Privacy Policy. Wikipedia® is a registered trademark of the Wikimedia
Foundation, Inc., a non-profit organization.

ANN - Wiki

Uploaded by

Copyright:

Available Formats

ANN - Wiki

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ANN - Wiki

Uploaded by

Copyright:

Available Formats

10/3/2018 Artificial neural network - Wikipedia

Artificial neural network

An ANN is based on a collection of connected units or nodes

An artificial neuron mimics the working of a

The network forms by connecting the output of

Components of an artificial neural network

an activation , depending on a discrete time parameter,

and an output function computing the output from the activation

Often the output function is simply the Identity function.

Connections, weights and biases

Neural networks as functions

This figure depicts such a decomposition of , with dependencies between variables

Choosing a cost function

Convergent recursive learning algorithm

Group method of data handling

Convolutional neural networks

Long short-term memory

Deep reservoir computing

Deep belief networks

Stacked (de-noising) auto-encoders

Deep stacking networks

which has a closed-form solution.

Tensor deep stacking networks

Compound hierarchical-deep models

Deep predictive coding networks

DPCNs can be extended to form a convolutional network.[173]

Networks with separate memory structures

LSTM-related differentiable memory structures

Neural Turing machines

Multilayer kernel machine

Neural architecture search

Generalization and statistics

By assigning a softmax activation function, a generalization of the logistic

The softmax activation function is:

Practical counterexamples to criticisms

Technology writer Roger Bridgman commented:

A single-layer feedforward artificial neural A two-layer feedforward artificial neural network.

An artificial neural network. An ANN dependency graph.

A single-layer feedforward artificial neural A two-layer feedforward artificial neural network

Parallel pipeline structure of CMAC neural

Retrieved from "https://en.wikipedia.org/w/index.php?title=Artificial_neural_network&oldid=862148531"

This page was last edited on 2 October 2018, at 13:46 (UTC).

You might also like