Artificial neural network

Artificial neural networks (ANNs), usually simply called neural networks (NNs) or neural nets,[1] are computing systems inspired by the biological neural networks that constitute animal brains.[2]

An artificial neural network is an interconnected group of nodes, inspired by a simplification of neurons in a brain. Here, each circular node represents an artificial neuron and an arrow represents a connection from the output of one artificial neuron to the input of another.

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 to other neurons. An artificial neuron receives signals then processes them and can signal neurons connected to it. The "signal" at a connection is a real number, and the output of each neuron is computed by some non-linear function of the sum of its inputs. The connections are called edges. 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. Neurons may have a threshold such that a signal is sent only if the aggregate signal crosses that threshold.

Typically, neurons are aggregated into layers. Different layers may perform different 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.

Learning Units edit

Models edit

Neuron and myelinated axon, with signal flow from inputs at dendrites to outputs at axon terminals

Hyperparameter edit

A hyperparameter is a constant parameter whose value is set before the learning process begins. The values of parameters are derived via learning. Examples of hyperparameters include learning rate, the number of hidden layers and batch size.[3] The values of some hyperparameters can be dependent on those of other hyperparameters. For example, the size of some layers can depend on the overall number of layers.

Learning edit

Learning is the adaptation of the network to better handle a task by considering sample observations. Learning involves adjusting the weights (and optional thresholds) of the network to improve the accuracy of the result. This is done by minimizing the observed errors. Learning is complete when examining additional observations does not usefully reduce the error rate. Even after learning, the error rate typically does not reach 0. If after learning, the error rate is too high, the network typically must be redesigned. Practically this is done by defining a cost function that is evaluated periodically during learning. As long as its output continues to decline, learning continues. The cost is frequently defined as a statistic whose value can only be approximated. The outputs are actually numbers, so when the error is low, the difference between the output (almost certainly a cat) and the correct answer (cat) is small. Learning attempts to reduce the total of the differences across the observations. Most learning models can be viewed as a straightforward application of optimization theory and statistical estimation.[4][5]

Learning rate edit

The learning rate defines the size of the corrective steps that the model takes to adjust for errors in each observation.[6] A high learning rate shortens the training time, but with lower ultimate accuracy, while a lower learning rate takes longer, but with the potential for greater accuracy. Optimizations such as Quickprop are primarily aimed at speeding up error minimization, while other improvements mainly try to increase reliability. In order to avoid oscillation inside the network such as alternating connection weights, and to improve the rate of convergence, refinements use an adaptive learning rate that increases or decreases as appropriate.[7] The concept of momentum allows the balance between the gradient and the previous change to be weighted such that the weight adjustment depends to some degree on the previous change. A momentum close to 0 emphasizes the gradient, while a value close to 1 emphasizes the last change.

Cost function edit

While it is possible to define a cost function ad hoc, frequently the choice is determined by the function's desirable properties (such as convexity) or because it arises from the model (e.g. in a probabilistic model the model's posterior probability can be used as an inverse cost).

Backpropagation edit

Backpropagation is a method used to adjust the connection weights to compensate for each error found during learning. The error amount is effectively divided among the connections. Technically, backprop calculates the gradient (the derivative) of the cost function associated with a given state with respect to the weights. The weight updates can be done via stochastic gradient descent or other methods, such as Extreme Learning Machines,[8] "No-prop" networks,[9] training without backtracking,[10] "weightless" networks,[11][12] and non-connectionist neural networks.[citation needed]

Learning paradigms edit

Machine learning is commonly separated into three main learning paradigms, supervised learning, unsupervised learning and reinforcement learning.[13] Each corresponds to a particular learning task.

Supervised learning edit

Supervised learning uses a set of paired inputs and desired outputs. The learning task is to produce the desired output for each input. In this case, the cost function is related to eliminating incorrect deductions.[14] A commonly used cost is the mean-squared error, which tries to minimize the average squared error between the network's output and the desired output. Tasks suited for supervised learning are pattern recognition (also known as classification) and regression (also known as function approximation). Supervised learning 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 edit

In unsupervised learning, input data is given along with the cost function, some 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 edit

In applications such as playing video games, an actor takes a string of actions, receiving a generally unpredictable response from the environment after each one. The goal is to win the game, i.e., generate the most positive (lowest cost) responses. In reinforcement learning, the aim is to weight the network (devise a policy) to perform actions that minimize long-term (expected cumulative) cost. 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) rules. The rules and the long-term cost usually only can be estimated. At any juncture, the agent decides whether to explore new actions to uncover their costs or to exploit prior learning to proceed more quickly.

Formally the environment is modeled as a Markov decision process (MDP) with states   and actions  . Because the state transitions are not known, probability distributions are used instead: the instantaneous cost distribution  , the observation distribution   and the transition distribution  , while a policy is defined as the conditional distribution over actions given the observations. Taken together, the two define a Markov chain (MC). The aim is to discover the lowest-cost MC.

ANNs serve as the learning component in such applications.[15][16] Dynamic programming coupled with ANNs (giving neurodynamic programming)[17] has been applied to problems such as those involved in vehicle routing,[18] video games, natural resource management[19][20] and medicine[21] because of ANNs ability to mitigate losses of accuracy even when reducing the discretization grid density for numerically approximating the solution of control problems. Tasks that fall within the paradigm of reinforcement learning are control problems, games and other sequential decision making tasks.

Self-learning edit

Self-learning in neural networks was introduced in 1982 along with a neural network capable of self-learning named Crossbar Adaptive Array (CAA).[22] It is a system with only one input, situation s, and only one output, action (or behavior) a. It has neither external advice input nor external reinforcement input from the environment. The CAA computes, in a crossbar fashion, both decisions about actions and emotions (feelings) about encountered situations. The system is driven by the interaction between cognition and emotion.[23] Given the memory matrix, W =||w(a,s)||, the crossbar self-learning algorithm in each iteration performs the following computation:

  In situation s perform action a;
  Receive consequence situation s';
  Compute emotion of being in consequence situation v(s');
  Update crossbar memory w'(a,s) = w(a,s) + v(s').

The backpropagated value (secondary reinforcement) is the emotion toward the consequence situation. The CAA exists in two environments, one is behavioral environment where it behaves, and the other is genetic environment, where from it initially and only once receives initial emotions about to be encountered situations in the behavioral environment. Having received the genome vector (species vector) from the genetic environment, the CAA will learn a goal-seeking behavior, in the behavioral environment that contains both desirable and undesirable situations.[24]

Neuroevolution edit

Neuroevolution can create neural network topologies and weights using evolutionary computation. It is competitive with sophisticated gradient descent approaches[citation needed]. One advantage of neuroevolution is that it may be less prone to get caught in "dead ends".[25]

Stochastic neural network edit

Stochastic neural networks originating from Sherrington–Kirkpatrick models are a type of artificial neural network built by introducing random variations into the network, either by giving the network's artificial neurons stochastic transfer functions, or by giving them stochastic weights. This makes them useful tools for optimization problems, since the random fluctuations help the network escape from local minima.[26] Stochastic neural networks trained using a Bayesian approach are known as Bayesian neural networks.[27]

Other edit

In a Bayesian framework, a distribution over the set of allowed models is chosen to minimize the cost. Evolutionary methods,[28] gene expression programming,[29] simulated annealing,[30] expectation-maximization, non-parametric methods and particle swarm optimization[31] are other learning algorithms. Convergent recursion is a learning algorithm for cerebellar model articulation controller (CMAC) neural networks.[32][33]

Modes edit

Two modes of learning are available: stochastic and batch. In stochastic learning, each input creates a weight adjustment. In batch learning weights are adjusted based on a batch of inputs, accumulating errors over the batch. Stochastic learning introduces "noise" into the process, using the local gradient calculated from one data point; this reduces the chance of the network getting stuck in local minima. However, batch learning typically yields a faster, more stable descent to a local minimum, since each update is performed in the direction of the batch's average error. A common compromise is to use "mini-batches", small batches with samples in each batch selected stochastically from the entire data set.

Types edit

ANNs have evolved into a broad family of techniques that have advanced the state of the art across multiple domains. The simplest 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 evolve via learning. 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.

Some of the main breakthroughs include: convolutional neural networks that have proven particularly successful in processing visual and other two-dimensional data;[34][35] long short-term memory avoid the vanishing gradient problem[36] and can handle signals that have a mix of low and high frequency components aiding large-vocabulary speech recognition,[37][38] text-to-speech synthesis,[39][40][41] and photo-real talking heads;[42] competitive networks such as generative adversarial networks in which multiple networks (of varying structure) compete with each other, on tasks such as winning a game[43] or on deceiving the opponent about the authenticity of an input.[44]

Network design edit

Neural architecture search (NAS) uses machine learning to automate ANN design. 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.[45] Available systems include AutoML and AutoKeras.[46]

Design issues include deciding the number, type and connectedness of network layers, as well as the size of each and the connection type (full, pooling, ...).

Hyperparameters must also be defined as part of the design (they are not learned), governing matters such as how many neurons are in each layer, learning rate, step, stride, depth, receptive field and padding (for CNNs), etc.[47]

Use edit

Template:Unreferenced section 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 are 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:[citation needed]

Applications edit

Because of their ability to reproduce and model nonlinear processes, artificial neural networks have found applications in many disciplines. Application areas include system identification and control (vehicle control, trajectory prediction,[49] process control, natural resource management), quantum chemistry,[50] general game playing,[51] pattern recognition (radar systems, face identification, signal classification,[52] 3D reconstruction,[53] object recognition and more), sensor data analysis,[54] sequence recognition (gesture, speech, handwritten and printed text recognition[55]), medical diagnosis, finance[56] (e.g. ex-ante models for specific financial long-run forecasts and artificial financial markets), data mining, visualization, machine translation, social network filtering[57] and e-mail spam filtering. ANNs have been used to diagnose several types of cancers[58][59] and to distinguish highly invasive cancer cell lines from less invasive lines using only cell shape information.[60][61]

ANNs have been used to accelerate reliability analysis of infrastructures subject to natural disasters[62][63] and to predict foundation settlements.[64] ANNs have also been used for building black-box models in geoscience: hydrology,[65][66] ocean modelling and coastal engineering,[67][68] and geomorphology.[69] ANNs have been employed in cybersecurity, with the objective to discriminate between legitimate activities and malicious ones. For example, machine learning has been used for classifying Android malware,[70] for identifying domains belonging to threat actors and for detecting URLs posing a security risk.[71] Research is underway on ANN systems designed for penetration testing, for detecting botnets,[72] credit cards frauds[73] and network intrusions.

ANNs have been proposed as a tool to solve partial differential equations in physics[74][75][76] and simulate the properties of many-body open quantum systems.[77][78][79][80] In brain research ANNs have studied short-term behavior of individual neurons,[81] the dynamics of neural circuitry arise from interactions between individual neurons and how behavior can arise from abstract neural modules that represent complete subsystems. Studies considered long-and short-term plasticity of neural systems and their relation to learning and memory from the individual neuron to the system level.

Theoretical properties edit

Computational power edit

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 power of a universal Turing machine,[82] 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.[83]

Capacity edit

A model's "capacity" property corresponds to its 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. Two notions of capacity are known by the community. The information capacity and the VC Dimension. The information capacity of a perceptron is intensively discussed in Sir David MacKay's book[84] which summarizes work by Thomas Cover.[85] The capacity of a network of standard neurons (not convolutional) can be derived by four rules[86] that derive from understanding a neuron as an electrical element. The information capacity captures the functions modelable by the network given any data as input. The second notion, is the VC dimension. VC Dimension uses the principles of measure theory and finds the maximum capacity under the best possible circumstances. This is, given input data in a specific form. As noted in,[84] the VC Dimension for arbitrary inputs is half the information capacity of a Perceptron. The VC Dimension for arbitrary points is sometimes referred to as Memory Capacity.[87]

Convergence edit

Models may not consistently converge on a single solution, firstly because 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.

Another issue worthy to mention is that training may cross some Saddle point which may lead the convergence to the wrong direction.

The convergence behavior of certain types of ANN architectures are more understood than others. When the width of network approaches to infinity, the ANN is well described by its first order Taylor expansion throughout training, and so inherits the convergence behavior of affine models.[88][89] Another example is when parameters are small, it is observed that ANNs often fits target functions from low to high frequencies. This behavior is referred to as the spectral bias, or frequency principle, of neural networks.[90][91][92][93] This phenomenon is the opposite to the behavior of some well studied iterative numerical schemes such as Jacobi method. Deeper neural networks have been observed to be more biased towards low frequency functions.[94]

Generalization and statistics edit

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 network capacity 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 to 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.

Confidence analysis of a neural network

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 network output, 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 a component-based network) for categorical target variables, the outputs can be interpreted as posterior probabilities. This is useful in classification as it gives a certainty measure on classifications.

The softmax activation function is:


Criticism edit

Training edit

A common criticism of neural networks, particularly in robotics, is that they require too much training for real-world operation.[95] 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, grouping examples in so-called mini-batches and/or introducing a recursive least squares algorithm for CMAC.[32]

Theory edit

A central claim[citation needed] of ANNs is that they embody new and powerful general principles for processing information. These principles are ill-defined. It is often claimedTemplate:By whom? 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. In 1997, 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".[96] One response to Dewdney is that neural networks handle many complex and diverse tasks, ranging from autonomously flying aircraft[97] to detecting credit card fraud to mastering the game of Go.

Technology writer Roger Bridgman commented:

Biological brains use both shallow and deep circuits as reported by brain anatomy,[99] displaying a wide variety of invariance. Weng[100] argued that the brain self-wires largely according to signal statistics and therefore, a serial cascade cannot catch all major statistical dependencies.

Hardware edit

Large and effective neural networks require considerable computing resources.[101] 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 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 require enormous CPU power and time.

Schmidhuber noted 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.[102] The use of accelerators such as FPGAs and GPUs can reduce training times from months to days.[101]

Neuromorphic engineering or a physical neural network addresses the hardware difficulty directly, by constructing non-von-Neumann chips to directly implement neural networks in circuitry. Another type of chip optimized for neural network processing is called a Tensor Processing Unit, or TPU.[103]

Practical counterexamples edit

Analyzing what has been learned by an ANN is much easier than analyzing 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.[104]

Hybrid approaches edit

Advocates of hybrid models (combining neural networks and symbolic approaches) say that such a mixture can better capture the mechanisms of the human mind.[105]

Gallery edit

See also edit


Notes edit

References edit

  1. Hardesty, Larry (14 April 2017). "Explained: Neural networks". MIT News Office. Retrieved 2 June 2022.
  2. Yang, Z.R.; Yang, Z. (2014). Comprehensive Biomedical Physics. Karolinska Institute, Stockholm, Sweden: Elsevier. p. 1. ISBN 978-0-444-53633-4. Retrieved 28 July 2022. 
  3. Lau, Suki (10 July 2017). "A Walkthrough of Convolutional Neural Network – Hyperparameter Tuning". Medium. Archived from the original on 4 February 2023. Retrieved 23 August 2019.
  4. Cite error: Invalid <ref> tag; no text was provided for refs named Zell1994ch5.2
  5. Kelleher, John D.; Mac Namee, Brian; D'Arcy, Aoife (2020). "7-8". Fundamentals of machine learning for predictive data analytics: algorithms, worked examples, and case studies (2nd ed.). Cambridge, MA. ISBN 978-0-262-36110-1. OCLC 1162184998. 
  6. Wei, Jiakai (2019-04-26). "Forget the Learning Rate, Decay Loss". arXiv:1905.00094 [cs.LG].
  7. Li, Y.; Fu, Y.; Li, H.; Zhang, S. W. (1 June 2009). The Improved Training Algorithm of Back Propagation Neural Network with Self-adaptive Learning Rate. 1. 73–76. doi:10.1109/CINC.2009.111. ISBN 978-0-7695-3645-3. 
  8. 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. 
  9. Widrow, Bernard (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. PMID 23140797. 
  10. Ollivier, Yann; Charpiat, Guillaume (2015). "Training recurrent networks without backtracking". arXiv:1507.07680 [cs.NE].
  11. Hinton, G. E. (2010). "A Practical Guide to Training Restricted Boltzmann Machines". Tech. Rep. UTML TR 2010-003. Archived from the original on 9 May 2021. Retrieved 27 June 2017. 
  12. ESANN. 2009.Template:Full citation needed
  13. Bernard, Etienne (2021). Introduction to Machine Learning. Wolfram Media Inc. p. 9. ISBN 978-1-579550-48-6. 
  14. Ojha, Varun Kumar; Abraham, Ajith; Snášel, Václav (1 April 2017). "Metaheuristic design of feedforward neural networks: A review of two decades of research". Engineering Applications of Artificial Intelligence 60: 97–116. doi:10.1016/j.engappai.2017.01.013. 
  15. 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. pp. 71–76. doi:10.1109/IJCNN.1991.155315. ISBN 0-7803-0164-1.
  16. 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. 
  17. Bertsekas, D.P.; Tsitsiklis, J.N. (1996). Neuro-dynamic programming. Athena Scientific. p. 512. ISBN 978-1-886529-10-6. Retrieved 17 June 2017. 
  18. 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. 
  19. de Rigo, D.; Rizzoli, A. E.; Soncini-Sessa, R.; Weber, E.; Zenesi, P. (2001). "Neuro-dynamic programming for the efficient management of reservoir networks". Proceedings of MODSIM 2001, International Congress on Modelling and Simulation. MODSIM 2001, International Congress on Modelling and Simulation. Canberra, Australia: Modelling and Simulation Society of Australia and New Zealand. doi:10.5281/zenodo.7481. ISBN 0-86740-525-2. Archived from the original on 7 August 2013. Retrieved 29 July 2013.
  20. 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. Vol. 1. La Jolla, California, USA: IEEE. pp. 7–14. doi:10.1109/CEC.2000.870269. ISBN 0-7803-6375-2.
  21. Deng, Geng; Ferris, M.C. (2008). Neuro-dynamic programming for fractionated radiotherapy planning. Springer Optimization and Its Applications. 12. pp. 47–70. doi:10.1007/978-0-387-73299-2_3. ISBN 978-0-387-73298-5. 
  22. Bozinovski, S. (1982). "A self-learning system using secondary reinforcement". In R. Trappl (ed.) Cybernetics and Systems Research: Proceedings of the Sixth European Meeting on Cybernetics and Systems Research. North Holland. pp. 397–402. ISBN 978-0-444-86488-8.
  23. Bozinovski, S. (2014) "Modeling mechanisms of cognition-emotion interaction in artificial neural networks, since 1981 Archived 23 March 2019 at the Wayback Machine." Procedia Computer Science p. 255-263
  24. Bozinovski, Stevo; Bozinovska, Liljana (2001). "Self-learning agents: A connectionist theory of emotion based on crossbar value judgment". Cybernetics and Systems 32 (6): 637–667. doi:10.1080/01969720118145. 
  25. "Artificial intelligence can 'evolve' to solve problems". Science | AAAS. 10 January 2018. Archived from the original on 9 December 2021. Retrieved 7 February 2018.
  26. Turchetti, Claudio (2004), Stochastic Models of Neural Networks, Frontiers in artificial intelligence and applications: Knowledge-based intelligent engineering systems, vol. 102, IOS Press, ISBN 9781586033880
  27. Jospin, Laurent Valentin; Laga, Hamid; Boussaid, Farid; Buntine, Wray; Bennamoun, Mohammed (2022). "Hands-On Bayesian Neural Networks—A Tutorial for Deep Learning Users". IEEE Computational Intelligence Magazine 17 (2): 29–48. doi:10.1109/mci.2022.3155327. ISSN 1556-603X. Archived from the original on 4 February 2023. Retrieved 19 November 2022. 
  28. 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". In Pavel Zítek (ed.). Proceedings of the 16th IFAC World Congress – IFAC-PapersOnLine. 16th IFAC World Congress. Vol. 16. Prague, Czech Republic: IFAC. pp. 7–12. doi:10.3182/20050703-6-CZ-1902.02172. hdl:11311/255236. ISBN 978-3-902661-75-3. Archived from the original on 26 April 2012. Retrieved 30 December 2011.
  29. Ferreira, C. (2006). "Designing Neural Networks Using Gene Expression Programming". In A. Abraham. Applied Soft Computing Technologies: The Challenge of Complexity. Springer-Verlag. pp. 517–536. Retrieved 8 October 2012. 
  30. Da, Y.; Xiurun, G. (July 2005). "An improved PSO-based ANN with simulated annealing technique". In T. Villmann (ed.). New Aspects in Neurocomputing: 11th European Symposium on Artificial Neural Networks. Vol. 63. Elsevier. pp. 527–533. doi:10.1016/j.neucom.2004.07.002. Archived from the original on 25 April 2012. Retrieved 30 December 2011.
  31. Wu, J.; Chen, E. (May 2009). "A Novel Nonparametric Regression Ensemble for Rainfall Forecasting Using Particle Swarm Optimization Technique Coupled with Artificial Neural Network". In Wang, H.; Shen, Y.; Huang, T.; Zeng, Z. (eds.). 6th International Symposium on Neural Networks, ISNN 2009. Lecture Notes in Computer Science. Vol. 5553. Springer. pp. 49–58. doi:10.1007/978-3-642-01513-7_6. ISBN 978-3-642-01215-0. Archived from the original on 31 December 2014. Retrieved 1 January 2012.
  32. 32.0 32.1 Ting Qin; Zonghai Chen; Haitao Zhang; Sifu Li; Wei Xiang; Ming Li (2004). "A learning algorithm of CMAC based on RLS". Neural Processing Letters 19 (1): 49–61. doi:10.1023/B:NEPL.0000016847.18175.60. Archived from the original on 14 April 2021. Retrieved 30 January 2019. 
  33. Ting Qin; Haitao Zhang; Zonghai Chen; Wei Xiang (2005). "Continuous CMAC-QRLS and its systolic array". Neural Processing Letters 22 (1): 1–16. doi:10.1007/s11063-004-2694-0. Archived from the original on 18 November 2018. Retrieved 30 January 2019. 
  34. "Backpropagation Applied to Handwritten Zip Code Recognition". Neural Computation 1 (4): 541–551. 1989. doi:10.1162/neco.1989.1.4.541. 
  35. Yann LeCun (2016). Slides on Deep Learning Online Archived 23 April 2016 at the Wayback Machine
  36. Hochreiter, Sepp; Schmidhuber, Jürgen (1 November 1997). "Long Short-Term Memory". Neural Computation 9 (8): 1735–1780. doi:10.1162/neco.1997.9.8.1735. ISSN 0899-7667. PMID 9377276. 
  37. Sak, Hasim; Senior, Andrew; Beaufays, Francoise (2014). "Long Short-Term Memory recurrent neural network architectures for large scale acoustic modeling" (PDF). Archived from the original (PDF) on 24 April 2018.
  38. Li, Xiangang; Wu, Xihong (15 October 2014). "Constructing Long Short-Term Memory based Deep Recurrent Neural Networks for Large Vocabulary Speech Recognition". arXiv:1410.4281 [cs.CL].
  39. Fan, Y.; Qian, Y.; Xie, F.; Soong, F. K. (2014). "TTS synthesis with bidirectional LSTM based Recurrent Neural Networks". Proceedings of the Annual Conference of the International Speech Communication Association, Interspeech: 1964–1968. Retrieved 13 June 2017. 
  40. Cite error: Invalid <ref> tag; no text was provided for refs named scholarpedia2
  41. Zen, Heiga; Sak, Hasim (2015). "Unidirectional Long Short-Term Memory Recurrent Neural Network with Recurrent Output Layer for Low-Latency Speech Synthesis" (PDF). ICASSP. pp. 4470–4474. Archived (PDF) from the original on 9 May 2021. Retrieved 27 June 2017.
  42. Fan, Bo; Wang, Lijuan; Soong, Frank K.; Xie, Lei (2015). "Photo-Real Talking Head with Deep Bidirectional LSTM". Proceedings of ICASSP. Archived from the original on 1 November 2017. Retrieved 27 June 2017. 
  43. Silver, David; Hubert, Thomas; Schrittwieser, Julian; Antonoglou, Ioannis; Lai, Matthew; Guez, Arthur; Lanctot, Marc; Sifre, Laurent; Kumaran, Dharshan; Graepel, Thore; Lillicrap, Timothy; Simonyan, Karen; Hassabis, Demis (5 December 2017). "Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm". arXiv:1712.01815 [cs.AI].
  44. Goodfellow, Ian; Pouget-Abadie, Jean; Mirza, Mehdi; Xu, Bing; Warde-Farley, David; Ozair, Sherjil; Courville, Aaron; Bengio, Yoshua (2014). Generative Adversarial Networks (PDF). Proceedings of the International Conference on Neural Information Processing Systems (NIPS 2014). pp. 2672–2680. Archived (PDF) from the original on 22 November 2019. Retrieved 20 August 2019.
  45. Zoph, Barret; Le, Quoc V. (4 November 2016). "Neural Architecture Search with Reinforcement Learning". arXiv:1611.01578 [cs.LG].
  46. Haifeng Jin; Qingquan Song; Xia Hu (2019). "Auto-keras: An efficient neural architecture search system". Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (ACM). Archived from the original on 21 August 2019. Retrieved 21 August 2019. 
  47. Claesen, Marc; De Moor, Bart (2015). "Hyperparameter Search in Machine Learning". arXiv:1502.02127 [cs.LG]. Bibcode2015arXiv150202127C
  48. Turek, Fred D. (March 2007). "Introduction to Neural Net Machine Vision". Vision Systems Design 12 (3). Archived from the original on 16 May 2013. Retrieved 5 March 2013. 
  49. Zissis, Dimitrios (October 2015). "A cloud based architecture capable of perceiving and predicting multiple vessel behaviour". Applied Soft Computing 35: 652–661. doi:10.1016/j.asoc.2015.07.002. Archived from the original on 26 July 2020. Retrieved 18 July 2019. 
  50. 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. doi:10.1063/1.3206326. PMID 19708729. 
  51. Silver, David (2016). "Mastering the game of Go with deep neural networks and tree search". Nature 529 (7587): 484–489. doi:10.1038/nature16961. PMID 26819042. Archived from the original on 23 November 2018. Retrieved 31 January 2019. 
  52. Sengupta, Nandini; Sahidullah, Md; Saha, Goutam (August 2016). "Lung sound classification using cepstral-based statistical features". Computers in Biology and Medicine 75 (1): 118–129. doi:10.1016/j.compbiomed.2016.05.013. PMID 27286184. 
  53. Choy, Christopher B., et al. "3d-r2n2: A unified approach for single and multi-view 3d object reconstruction Archived 26 July 2020 at the Wayback Machine." European conference on computer vision. Springer, Cham, 2016.
  54. Gessler, Josef (August 2021). "Sensor for food analysis applying impedance spectroscopy and artificial neural networks". RiuNet UPV (1): 8–12. Archived from the original on 21 October 2021. Retrieved 21 October 2021. 
  55. Maitra, D. S.; Bhattacharya, U.; Parui, S. K. (August 2015). "CNN based common approach to handwritten character recognition of multiple scripts". 2015 13th International Conference on Document Analysis and Recognition (ICDAR): 1021–1025. doi:10.1109/ICDAR.2015.7333916. ISBN 978-1-4799-1805-8. 
  56. French, Jordan (2016). "The time traveller's CAPM". Investment Analysts Journal 46 (2): 81–96. doi:10.1080/10293523.2016.1255469. 
  57. Schechner, Sam (15 June 2017). "Facebook Boosts A.I. to Block Terrorist Propaganda". The Wall Street Journal. ISSN 0099-9660. Retrieved 16 June 2017.
  58. Ganesan, N (2010). "Application of Neural Networks in Diagnosing Cancer Disease Using Demographic Data". International Journal of Computer Applications 1 (26): 81–97. doi:10.5120/476-783. 
  59. Bottaci, Leonardo (1997). "Artificial Neural Networks Applied to Outcome Prediction for Colorectal Cancer Patients in Separate Institutions". Lancet (The Lancet) 350 (9076): 469–72. doi:10.1016/S0140-6736(96)11196-X. PMID 9274582. Archived from the original on 23 November 2018. Retrieved 2 May 2012. 
  60. Alizadeh, Elaheh; Lyons, Samanthe M; Castle, Jordan M; Prasad, Ashok (2016). "Measuring systematic changes in invasive cancer cell shape using Zernike moments". Integrative Biology 8 (11): 1183–1193. doi:10.1039/C6IB00100A. PMID 27735002. 
  61. Lyons, Samanthe (2016). "Changes in cell shape are correlated with metastatic potential in murine". Biology Open 5 (3): 289–299. doi:10.1242/bio.013409. PMID 26873952. PMC 4810736. // 
  62. Nabian, Mohammad Amin; Meidani, Hadi (28 August 2017). "Deep Learning for Accelerated Reliability Analysis of Infrastructure Networks". Computer-Aided Civil and Infrastructure Engineering 33 (6): 443–458. doi:10.1111/mice.12359. 
  63. Nabian, Mohammad Amin; Meidani, Hadi (2018). "Accelerating Stochastic Assessment of Post-Earthquake Transportation Network Connectivity via Machine-Learning-Based Surrogates". Transportation Research Board 97th Annual Meeting. Archived from the original on 9 March 2018. Retrieved 14 March 2018. 
  64. Díaz, E.; Brotons, V.; Tomás, R. (September 2018). "Use of artificial neural networks to predict 3-D elastic settlement of foundations on soils with inclined bedrock". Soils and Foundations 58 (6): 1414–1422. doi:10.1016/j.sandf.2018.08.001. ISSN 0038-0806. 
  65. Govindaraju, Rao S. (1 April 2000). "Artificial Neural Networks in Hydrology. I: Preliminary Concepts". Journal of Hydrologic Engineering 5 (2): 115–123. doi:10.1061/(ASCE)1084-0699(2000)5:2(115). 
  66. Govindaraju, Rao S. (1 April 2000). "Artificial Neural Networks in Hydrology. II: Hydrologic Applications". Journal of Hydrologic Engineering 5 (2): 124–137. doi:10.1061/(ASCE)1084-0699(2000)5:2(124). 
  67. Peres, D. J.; Iuppa, C.; Cavallaro, L.; Cancelliere, A.; Foti, E. (1 October 2015). "Significant wave height record extension by neural networks and reanalysis wind data". Ocean Modelling 94: 128–140. doi:10.1016/j.ocemod.2015.08.002. 
  68. Dwarakish, G. S.; Rakshith, Shetty; Natesan, Usha (2013). "Review on Applications of Neural Network in Coastal Engineering". Artificial Intelligent Systems and Machine Learning 5 (7): 324–331. Archived from the original on 15 August 2017. Retrieved 5 July 2017. 
  69. Ermini, Leonardo; Catani, Filippo; Casagli, Nicola (1 March 2005). "Artificial Neural Networks applied to landslide susceptibility assessment". Geomorphology. Geomorphological hazard and human impact in mountain environments 66 (1): 327–343. doi:10.1016/j.geomorph.2004.09.025. 
  70. Nix, R.; Zhang, J. (May 2017). "Classification of Android apps and malware using deep neural networks". 2017 International Joint Conference on Neural Networks (IJCNN): 1871–1878. doi:10.1109/IJCNN.2017.7966078. ISBN 978-1-5090-6182-2. 
  71. "Detecting Malicious URLs". The systems and networking group at UCSD. Archived from the original on 14 July 2019. Retrieved 15 February 2019.
  72. Homayoun, Sajad; Ahmadzadeh, Marzieh; Hashemi, Sattar; Dehghantanha, Ali; Khayami, Raouf (2018), Dehghantanha, Ali; Conti, Mauro; Dargahi, Tooska (eds.), "BoTShark: A Deep Learning Approach for Botnet Traffic Detection", Cyber Threat Intelligence, Advances in Information Security, Springer International Publishing, pp. 137–153, doi:10.1007/978-3-319-73951-9_7, ISBN 978-3-319-73951-9
  73. Ghosh; Reilly (January 1994). "Credit card fraud detection with a neural-network". 1994 Proceedings of the Twenty-Seventh Hawaii International Conference on System Sciences 3: 621–630. doi:10.1109/HICSS.1994.323314. ISBN 978-0-8186-5090-1. 
  74. Ananthaswamy, Anil (19 April 2021). "Latest Neural Nets Solve World's Hardest Equations Faster Than Ever Before". Quanta Magazine. Retrieved 12 May 2021.
  75. "AI has cracked a key mathematical puzzle for understanding our world". MIT Technology Review. Retrieved 19 November 2020.
  76. "Caltech Open-Sources AI for Solving Partial Differential Equations". InfoQ. Archived from the original on 25 January 2021. Retrieved 20 January 2021.
  77. Nagy, Alexandra (28 June 2019). "Variational Quantum Monte Carlo Method with a Neural-Network Ansatz for Open Quantum Systems". Physical Review Letters 122 (25): 250501. doi:10.1103/PhysRevLett.122.250501. PMID 31347886. 
  78. Yoshioka, Nobuyuki; Hamazaki, Ryusuke (28 June 2019). "Constructing neural stationary states for open quantum many-body systems". Physical Review B 99 (21): 214306. doi:10.1103/PhysRevB.99.214306. 
  79. Hartmann, Michael J.; Carleo, Giuseppe (28 June 2019). "Neural-Network Approach to Dissipative Quantum Many-Body Dynamics". Physical Review Letters 122 (25): 250502. doi:10.1103/PhysRevLett.122.250502. PMID 31347862. 
  80. Vicentini, Filippo; Biella, Alberto; Regnault, Nicolas; Ciuti, Cristiano (28 June 2019). "Variational Neural-Network Ansatz for Steady States in Open Quantum Systems". Physical Review Letters 122 (25): 250503. doi:10.1103/PhysRevLett.122.250503. PMID 31347877. 
  81. Forrest MD (April 2015). "Simulation of alcohol action upon a detailed Purkinje neuron model and a simpler surrogate model that runs >400 times faster". BMC Neuroscience 16 (27): 27. doi:10.1186/s12868-015-0162-6. PMID 25928094. PMC 4417229. // 
  82. Siegelmann, H.T.; Sontag, E.D. (1991). "Turing computability with neural nets". Appl. Math. Lett. 4 (6): 77–80. doi:10.1016/0893-9659(91)90080-F. 
  83. Balcázar, José (Jul 1997). "Computational Power of Neural Networks: A Kolmogorov Complexity Characterization". IEEE Transactions on Information Theory 43 (4): 1175–1183. doi:10.1109/18.605580. 
  84. 84.0 84.1 MacKay, David, J.C. (2003). Information Theory, Inference, and Learning Algorithms. Cambridge University Press. ISBN 978-0-521-64298-9. Retrieved 11 June 2016. 
  85. Cover, Thomas (1965). "Geometrical and Statistical Properties of Systems of Linear Inequalities with Applications in Pattern Recognition". IEEE Transactions on Electronic Computers (IEEE) EC-14 (3): 326–334. doi:10.1109/PGEC.1965.264137. Archived from the original on 5 March 2016. Retrieved 10 March 2020. 
  86. Gerald, Friedland (2019). "Reproducibility and Experimental Design for Machine Learning on Audio and Multimedia Data". MM '19: Proceedings of the 27th ACM International Conference on Multimedia (ACM): 2709–2710. doi:10.1145/3343031.3350545. ISBN 978-1-4503-6889-6. 
  87. "Stop tinkering, start measuring! Predictable experimental design of Neural Network experiments". The Tensorflow Meter. Archived from the original on 18 April 2022. Retrieved 10 March 2020.
  88. Lee, Jaehoon; Xiao, Lechao; Schoenholz, Samuel S.; Bahri, Yasaman; Novak, Roman; Sohl-Dickstein, Jascha; Pennington, Jeffrey (2020). "Wide neural networks of any depth evolve as linear models under gradient descent". Journal of Statistical Mechanics: Theory and Experiment 2020 (12): 124002. doi:10.1088/1742-5468/abc62b. 
  89. Arthur Jacot; Franck Gabriel; Clement Hongler (2018). Neural Tangent Kernel: Convergence and Generalization in Neural Networks (PDF). 32nd Conference on Neural Information Processing Systems (NeurIPS 2018), Montreal, Canada. Archived (PDF) from the original on 22 June 2022. Retrieved 4 June 2022.
  90. "Training Behavior of Deep Neural Network in Frequency Domain". Neural Information Processing. ICONIP 2019.. Lecture Notes in Computer Science. 11953. Springer, Cham. 2019. pp. 264–274. doi:10.1007/978-3-030-36708-4_22. ISBN 978-3-030-36707-7. 
  91. Nasim Rahaman; Aristide Baratin; Devansh Arpit; Felix Draxler; Min Lin; Fred Hamprecht; Yoshua Bengio; Aaron Courville (2019). "On the Spectral Bias of Neural Networks". Proceedings of the 36th International Conference on Machine Learning 97: 5301–5310. Archived from the original on 22 October 2022. Retrieved 4 June 2022. 
  92. Zhi-Qin John Xu; Yaoyu Zhang; Tao Luo; Yanyang Xiao; Zheng Ma (2020). "Frequency Principle: Fourier Analysis Sheds Light on Deep Neural Networks". Communications in Computational Physics 28 (5): 1746–1767. doi:10.4208/cicp.OA-2020-0085. 
  93. Tao Luo; Zheng Ma; Zhi-Qin John Xu; Yaoyu Zhang (2019). "Theory of the Frequency Principle for General Deep Neural Networks". arXiv:1906.09235 [cs.LG].
  94. Xu, Zhiqin John; Zhou, Hanxu (2021-05-18). "Deep Frequency Principle Towards Understanding Why Deeper Learning Is Faster". Proceedings of the AAAI Conference on Artificial Intelligence 35 (12): 10541–10550. doi:10.1609/aaai.v35i12.17261. ISSN 2374-3468. Archived from the original on 5 October 2021. Retrieved 5 October 2021. 
  95. Parisi, German I.; Kemker, Ronald; Part, Jose L.; Kanan, Christopher; Wermter, Stefan (2019-05-01). "Continual lifelong learning with neural networks: A review". Neural Networks 113: 54–71. doi:10.1016/j.neunet.2019.01.012. ISSN 0893-6080. 
  96. Dewdney, A. K. (1 April 1997). Yes, we have no neutrons: an eye-opening tour through the twists and turns of bad science. Wiley. pp. 82. ISBN 978-0-471-10806-1. 
  97. NASA – Dryden Flight Research Center – News Room: News Releases: NASA NEURAL NETWORK PROJECT PASSES MILESTONE Archived 2 April 2010 at the Wayback Machine. Retrieved on 20 November 2013.
  98. "Roger Bridgman's defence of neural networks". Archived from the original on 19 March 2012. Retrieved 12 July 2010.
  99. D. J. Felleman and D. C. Van Essen, "Distributed hierarchical processing in the primate cerebral cortex," Cerebral Cortex, 1, pp. 1–47, 1991.
  100. J. Weng, "Natural and Artificial Intelligence: Introduction to Computational Brain-Mind," BMI Press, ISBN 978-0-9858757-2-5, 2012.
  101. 101.0 101.1 Edwards, Chris (25 June 2015). "Growing pains for deep learning". Communications of the ACM 58 (7): 14–16. doi:10.1145/2771283. 
  102. Cite error: Invalid <ref> tag; no text was provided for refs named SCHIDHUB2
  103. Cade Metz (18 May 2016). "Google Built Its Very Own Chips to Power Its AI Bots". Wired. Archived from the original on 13 January 2018. Retrieved 5 March 2017.
  104. "Scaling Learning Algorithms towards AI" (PDF). Archived (PDF) from the original on 12 August 2022. Retrieved 6 July 2022.
  105. Tahmasebi; Hezarkhani (2012). "A hybrid neural networks-fuzzy logic-genetic algorithm for grade estimation". Computers & Geosciences 42: 18–27. doi:10.1016/j.cageo.2012.02.004. PMID 25540468. PMC 4268588. // 

Bibliography edit

Template:Complex systems topics Template:Control theory Template:Differentiable computing Template:Neuroscience Template:Self-driving cars and enabling technologies

Page Information edit

This page was based on the following wikipedia-source page: