We are proudly a Ukrainian website. Our country was attacked by Russian Armed Forces on Feb. 24, 2022.
You can support the Ukrainian Army by following the link: https://u24.gov.ua/.
Even the smallest donation is hugely appreciated!
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the lists of relevant articles, books, theses, conference reports, and other scholarly sources on the topic 'Neural networks.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Navghare, Tukaram, Aniket Muley, and Vinayak Jadhav. "Siamese Neural Networks for Kinship Prediction: A Deep Convolutional Neural Network Approach." Indian Journal Of Science And Technology 17, no. 4 (January 26, 2024): 352–58. http://dx.doi.org/10.17485/ijst/v17i4.3018.
N, Vikram. "Artificial Neural Networks." International Journal of Research Publication and Reviews 4, no. 4 (April 23, 2023): 4308–9. http://dx.doi.org/10.55248/gengpi.4.423.37858.
O. H. Abdelwahed, O. H. Abdelwahed, and M. El-Sayed Wahed. "Optimizing Single Layer Cellular Neural Network Simulator using Simulated Annealing Technique with Neural Networks." Indian Journal of Applied Research 3, no. 6 (October 1, 2011): 91–94. http://dx.doi.org/10.15373/2249555x/june2013/31.
<p>As a vector transmission network structure, the capsule neural network has been one of the research hotspots in deep learning since it was proposed in 2017. In this paper, the latest research progress of capsule networks is analyzed and summarized. Firstly, we summarize the shortcomings of convolutional neural networks and introduce the basic concept of capsule network. Secondly, we analyze and summarize the improvements in the dynamic routing mechanism and network structure of the capsule network in recent years and the combination of the capsule network with other network structures. Finally, we compile the applications of capsule network in many fields, including computer vision, natural language, and speech processing. Our purpose in writing this article is to provide methods and means that can be used for reference in the research and practical applications of capsule networks.</p> <p> </p>
5
Perfetti, R. "A neural network to design neural networks." IEEE Transactions on Circuits and Systems 38, no. 9 (1991): 1099–103. http://dx.doi.org/10.1109/31.83884.
To posses relevant information is an inevitable condition for successful enterprising in modern business. Information could be parted to data and knowledge. How to gather, store and retrieve data is studied in database theory. In the knowledge engineering, there is in the centre of interest the knowledge and methods of its formalization and gaining are studied. Knowledge could be gained from experts, specialists in the area of interest, or it can be gained by induction from sets of data. Automatic induction of knowledge from data sets, usually stored in large databases, is called data mining. Classical methods of gaining knowledge from data sets are statistical methods. In data mining, new methods besides statistical are used. These new methods have their origin in artificial intelligence. They look for unknown and unexpected relations, which can be uncovered by exploring of data in database. In the article, a utilization of modern methods of data mining is described and especially the methods based on neural networks theory are pursued. The advantages and drawbacks of applications of multiplayer feed forward neural networks and Kohonen’s self-organizing maps are discussed. Kohonen’s self-organizing map is the most promising neural data-mining algorithm regarding its capability to visualize high-dimensional data.
7
J, Joselin, Dinesh T, and Ashiq M. "A Review on Neural Networks." International Journal of Trend in Scientific Research and Development Volume-2, Issue-6 (October 31, 2018): 565–69. http://dx.doi.org/10.31142/ijtsrd18461.
Ziroyan, M. A., E. A. Tusova, A. S. Hovakimian, and S. G. Sargsyan. "Neural networks apparatus in biometrics." Contemporary problems of social work 1, no. 2 (June 30, 2015): 129–37. http://dx.doi.org/10.17922/2412-5466-2015-1-2-129-137.
Marton, Sascha, Stefan Lüdtke, and Christian Bartelt. "Explanations for Neural Networks by Neural Networks." Applied Sciences 12, no. 3 (January 18, 2022): 980. http://dx.doi.org/10.3390/app12030980.
Understanding the function learned by a neural network is crucial in many domains, e.g., to detect a model’s adaption to concept drift in online learning. Existing global surrogate model approaches generate explanations by maximizing the fidelity between the neural network and a surrogate model on a sample-basis, which can be very time-consuming. Therefore, these approaches are not applicable in scenarios where timely or frequent explanations are required. In this paper, we introduce a real-time approach for generating a symbolic representation of the function learned by a neural network. Our idea is to generate explanations via another neural network (called the Interpretation Network, or I-Net), which maps network parameters to a symbolic representation of the network function. We show that the training of an I-Net for a family of functions can be performed up-front and subsequent generation of an explanation only requires querying the I-Net once, which is computationally very efficient and does not require training data. We empirically evaluate our approach for the case of low-order polynomials as explanations, and show that it achieves competitive results for various data and function complexities. To the best of our knowledge, this is the first approach that attempts to learn mapping from neural networks to symbolic representations.
10
Wang, Jun. "Artificial neural networks versus natural neural networks." Decision Support Systems 11, no. 5 (June 1994): 415–29. http://dx.doi.org/10.1016/0167-9236(94)90016-7.
Dissertations / Theses on the topic "Neural networks":
1
Patterson, Raymond A. "Hybrid Neural networks and network design." Connect to resource, 1995. http://rave.ohiolink.edu/etdc/view.cgi?acc%5Fnum=osu1262707683.
Chambers, Mark Andrew. "Queuing network construction using artificial neural networks /." The Ohio State University, 2000. http://rave.ohiolink.edu/etdc/view?acc_num=osu1488193665234291.
Dunn, Nathan A. "A Novel Neural Network Analysis Method Applied to Biological Neural Networks." Thesis, view abstract or download file of text, 2006. http://proquest.umi.com/pqdweb?did=1251892251&sid=2&Fmt=2&clientId=11238&RQT=309&VName=PQD.
Thesis (Ph. D.)--University of Oregon, 2006. Typescript. Includes vita and abstract. Includes bibliographical references (leaves 122- 131). Also available for download via the World Wide Web; free to University of Oregon users.
5
Dong, Yue. "Higher Order Neural Networks and Neural Networks for Stream Learning." Thesis, Université d'Ottawa / University of Ottawa, 2017. http://hdl.handle.net/10393/35731.
The goal of this thesis is to explore some variations of neural networks. The thesis is mainly split into two parts: a variation of the shaping functions in neural networks and a variation of learning rules in neural networks.
In the first part, we mainly investigate polynomial perceptrons - a perceptron with a polynomial shaping function instead of a linear one. We prove the polynomial perceptron convergence theorem and illustrate the notion by showing that a higher order perceptron can learn the XOR function through empirical experiments with implementation. In the second part, we propose three models (SMLP, SA, SA2) for stream learning and anomaly detection in streams. The main technique allowing these models to perform at a level comparable to the state-of-the-art algorithms in stream learning is the learning rule used. We employ mini-batch gradient descent algorithm and stochastic gradient descent algorithm to speed up the models. In addition, the use of parallel processing with multi-threads makes the proposed methods highly efficient in dealing with streaming data. Our analysis shows that all models have linear runtime and constant memory requirement. We also demonstrate empirically that the proposed methods feature high detection rate, low false alarm rate, and fast response.
The paper on the first two models (SMLP, SA) is published in the 29th Canadian AI Conference and won the best paper award. The invited journal paper on the third model (SA2) for Computational Intelligence is under peer review.
6
Xu, Shuxiang, University of Western Sydney, and of Informatics Science and Technology Faculty. "Neuron-adaptive neural network models and applications." THESIS_FIST_XXX_Xu_S.xml, 1999. http://handle.uws.edu.au:8081/1959.7/275.
Artificial Neural Networks have been widely probed by worldwide researchers to cope with the problems such as function approximation and data simulation. This thesis deals with Feed-forward Neural Networks (FNN's) with a new neuron activation function called Neuron-adaptive Activation Function (NAF), and Feed-forward Higher Order Neural Networks (HONN's) with this new neuron activation function. We have designed a new neural network model, the Neuron-Adaptive Neural Network (NANN), and mathematically proved that one NANN can approximate any piecewise continuous function to any desired accuracy. In the neural network literature only Zhang proved the universal approximation ability of FNN Group to any piecewise continuous function. Next, we have developed the approximation properties of Neuron Adaptive Higher Order Neural Networks (NAHONN's), a combination of HONN's and NAF, to any continuous function, functional and operator. Finally, we have created a software program called MASFinance which runs on the Solaris system for the approximation of continuous or discontinuous functions, and for the simulation of any continuous or discontinuous data (especially financial data). Our work distinguishes itself from previous work in the following ways: we use a new neuron-adaptive activation function, while the neuron activation functions in most existing work are all fixed and can't be tuned to adapt to different approximation problems; we only use on NANN to approximate any piecewise continuous function, while a neural network group must be utilised in previous research; we combine HONN's with NAF and investigate its approximation properties to any continuous function, functional, and operator; we present a new software program, MASFinance, for function approximation and data simulation. Experiments running MASFinance indicate that the proposed NANN's present several advantages over traditional neuron-fixed networks (such as greatly reduced network size, faster learning, and lessened simulation errors), and that the suggested NANN's can effectively approximate piecewise continuous functions better than neural networks groups. Experiments also indicate that NANN's are especially suitable for data simulation Doctor of Philosophy (PhD)
7
Allen, T. J. "Optoelectronic neural networks." Thesis, University of Nottingham, 1996. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.362900.
Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018. This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. Cataloged from student-submitted PDF version of thesis. Includes bibliographical references (pages 65-68). Bus schedules are unreliable, leaving passengers waiting and increasing commute times. This problem can be solved by modeling the traffic network, and delivering predicted arrival times to passengers. Research attempts to model traffic networks use historical, statistical and learning based models, with learning based models achieving the best results. This research compares several neural network architectures trained on historical data from Boston buses. Three models are trained: multilayer perceptron, convolutional neural network and recurrent neural network. Recurrent neural networks show the best performance when compared to feed forward models. This indicates that neural time series models are effective at modeling bus networks. The large amount of data available for training bus network models and the effectiveness of large neural networks at modeling this data show that great progress can be made in improving commutes for passengers. by Cooper Stokes Sloan. M. Eng.
9
Boychenko, I. V., and G. I. Litvinenko. "Artificial neural networks." Thesis, Вид-во СумДУ, 2009. http://essuir.sumdu.edu.ua/handle/123456789/17044.
Landry, Kenneth D. "Evolutionary neural networks." Thesis, Virginia Polytechnic Institute and State University, 1988. http://hdl.handle.net/10919/51904.
To create neural networks that work, one needs to specify a structure and the interconnection weights between each pair of connected computing elements. The structure of a network can be selected by the designer depending on the application, although the selection of interconnection weights is a much larger problem. Algorithms have been developed to alter the weights slightly in order to produce the desired results. Learning algorithms such as Hebb's rule, the Delta rule and error propagation have been used, with success, to learn the appropriate weights. The major objection to this class of algorithms is that one cannot specify what is not desired in the network in addition to what is desired. An alternate method to learning the correct interconnection weights is to evolve a network in an environment that rewards "good” behavior and punishes "bad" behavior, This technique allows interesting networks to appear which otherwise may not be discovered by other methods of learning. In order to teach a network the correct weights, this approach simply needs a direction where an acceptable solution can be obtained rather than a complete answer to the problem. Master of Science
Abdi, Hervé, Dominique Valentin, and Betty Edelman. Neural Networks. 2455 Teller Road, Thousand Oaks California 91320 United States of America: SAGE Publications, Inc., 1999. http://dx.doi.org/10.4135/9781412985277.
Müller, Berndt, and Joachim Reinhardt. Neural Networks. Berlin, Heidelberg: Springer Berlin Heidelberg, 1990. http://dx.doi.org/10.1007/978-3-642-97239-3.
Müller, Berndt, Joachim Reinhardt, and Michael T. Strickland. Neural Networks. Berlin, Heidelberg: Springer Berlin Heidelberg, 1995. http://dx.doi.org/10.1007/978-3-642-57760-4.
Almeida, Luis B., and Christian J. Wellekens, eds. Neural Networks. Berlin, Heidelberg: Springer Berlin Heidelberg, 1990. http://dx.doi.org/10.1007/3-540-52255-7.
Wüthrich, Mario V., and Michael Merz. "Recurrent Neural Networks." In Springer Actuarial, 381–406. Cham: Springer International Publishing, 2022. http://dx.doi.org/10.1007/978-3-031-12409-9_8.
AbstractThis chapter considers recurrent neural (RN) networks. These are special network architectures that are useful for time-series modeling, e.g., applied to time-series forecasting. We study the most popular RN networks which are the long short-term memory (LSTM) networks and the gated recurrent unit (GRU) networks. We apply these networks to mortality forecasting.
3
Müller, Berndt, Joachim Reinhardt, and Michael T. Strickland. "Cybernetic Networks." In Neural Networks, 46–51. Berlin, Heidelberg: Springer Berlin Heidelberg, 1995. http://dx.doi.org/10.1007/978-3-642-57760-4_5.
Müller, Berndt, and Joachim Reinhardt. "Cybernetic Networks." In Neural Networks, 45–50. Berlin, Heidelberg: Springer Berlin Heidelberg, 1990. http://dx.doi.org/10.1007/978-3-642-97239-3_5.
Bungay, Henry R. "Neural Networks." In Environmental Systems Engineering, 121–38. Boston, MA: Springer US, 1998. http://dx.doi.org/10.1007/978-1-4615-5507-0_7.
Cios, Krzysztof J., Witold Pedrycz, and Roman W. Swiniarski. "Neural Networks." In Data Mining Methods for Knowledge Discovery, 309–74. Boston, MA: Springer US, 1998. http://dx.doi.org/10.1007/978-1-4615-5589-6_7.
Yang, Zhun, Adam Ishay, and Joohyung Lee. "NeurASP: Embracing Neural Networks into Answer Set Programming." In Twenty-Ninth International Joint Conference on Artificial Intelligence and Seventeenth Pacific Rim International Conference on Artificial Intelligence {IJCAI-PRICAI-20}. California: International Joint Conferences on Artificial Intelligence Organization, 2020. http://dx.doi.org/10.24963/ijcai.2020/243.
We present NeurASP, a simple extension of answer set programs by embracing neural networks. By treating the neural network output as the probability distribution over atomic facts in answer set programs, NeurASP provides a simple and effective way to integrate sub-symbolic and symbolic computation. We demonstrate how NeurASP can make use of a pre-trained neural network in symbolic computation and how it can improve the neural network's perception result by applying symbolic reasoning in answer set programming. Also, NeurASP can make use of ASP rules to train a neural network better so that a neural network not only learns from implicit correlations from the data but also from the explicit complex semantic constraints expressed by the rules.
2
Shi, Weijia, Andy Shih, Adnan Darwiche, and Arthur Choi. "On Tractable Representations of Binary Neural Networks." In 17th International Conference on Principles of Knowledge Representation and Reasoning {KR-2020}. California: International Joint Conferences on Artificial Intelligence Organization, 2020. http://dx.doi.org/10.24963/kr.2020/91.
We consider the compilation of a binary neural network’s decision function into tractable representations such as Ordered Binary Decision Diagrams (OBDDs) and Sentential Decision Diagrams (SDDs). Obtaining this function as an OBDD/SDD facilitates the explanation and formal verification of a neural network’s behavior. First, we consider the task of verifying the robustness of a neural network, and show how we can compute the expected robustness of a neural network, given an OBDD/SDD representation of it. Next, we consider a more efficient approach for compiling neural networks, based on a pseudo-polynomial time algorithm for compiling a neuron. We then provide a case study in a handwritten digits dataset, highlighting how two neural networks trained from the same dataset can have very high accuracies, yet have very different levels of robustness. Finally, in experiments, we show that it is feasible to obtain compact representations of neural networks as SDDs.
3
Najmon, Joel C., and Andres Tovar. "Comparing Derivatives of Neural Networks for Regression." In ASME 2023 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. American Society of Mechanical Engineers, 2023. http://dx.doi.org/10.1115/detc2023-117571.
Abstract In the past decades, neural networks have rapidly grown in popularity as a way to model complex non-linear relationships. The computational efficiently and flexibility of neural networks has made them popular for machine learning-based optimization methods. As such the derivative of a neural network’s output is required for gradient-based optimization algorithms. Recently, there have been several works towards improving derivatives of neural network targets, however there is yet to be done a comparative study on the different derivation methods for the derivative of a neural network’s targets with respect to its input features. Consequently, this paper’s objective is to implement and compare common methods for obtaining or approximating the derivative of neural network targets with respect to their inputs. The methods studied include analytical derivatives, finite differences, complex step approximation, and automatic differentiation. The methods are tested by training deep multilayer perceptrons for regression with several analytical functions. The derivatives of the neural network-derived methods are evaluated against the exact derivative of the test functions. Results show that all of the derivation methods provide the same derivative approximation to near working precision of the computer. Implementation of the study is done using the TensorFlow library in a provided Python code.
4
Ozcan, Neyir. "New results for global stability of neutral-type delayed neural networks." In The 11th International Conference on Integrated Modeling and Analysis in Applied Control and Automation. CAL-TEK srl, 2018. http://dx.doi.org/10.46354/i3m.2018.imaaca.004.
"This paper deals with the stability analysis of the class of neutral-type neural networks with constant time delay. By using a suitable Lyapunov functional, some delay independent sufficient conditions are derived, which ensure the global asymptotic stability of the equilibrium point for this this class of neutral-type neural networks with time delays with respect to the Lipschitz activation functions. The presented stability results rely on checking some certain properties of matrices. Therefore, it is easy to verify the validation of the constraint conditions on the network parameters of neural system by simply using some basic information of the matrix theory."
5
Benmaghnia, Hanane, Matthieu Martel, and Yassamine Seladji. "Fixed-Point Code Synthesis for Neural Networks." In 6th International Conference on Artificial Intelligence, Soft Computing and Applications (AISCA 2022). Academy and Industry Research Collaboration Center (AIRCC), 2022. http://dx.doi.org/10.5121/csit.2022.120202.
Over the last few years, neural networks have started penetrating safety critical systems to take decisions in robots, rockets, autonomous driving car, etc. A problem is that these critical systems often have limited computing resources. Often, they use the fixed-point arithmetic for its many advantages (rapidity, compatibility with small memory devices.) In this article, a new technique is introduced to tune the formats (precision) of already trained neural networks using fixed-point arithmetic, which can be implemented using integer operations only. The new optimized neural network computes the output with fixed-point numbers without modifying the accuracy up to a threshold fixed by the user. A fixed-point code is synthesized for the new optimized neural network ensuring the respect of the threshold for any input vector belonging the range [xmin, xmax] determined during the analysis. From a technical point of view, we do a preliminary analysis of our floating neural network to determine the worst cases, then we generate a system of linear constraints among integer variables that we can solve by linear programming. The solution of this system is the new fixed-point format of each neuron. The experimental results obtained show the efficiency of our method which can ensure that the new fixed-point neural network has the same behavior as the initial floating-point neural network.
6
Adewusi, S. A., and B. O. Al-Bedoor. "Detection of Propagating Cracks in Rotors Using Neural Networks." In ASME 2002 Pressure Vessels and Piping Conference. ASMEDC, 2002. http://dx.doi.org/10.1115/pvp2002-1518.
This paper presents the application of neural networks for rotor cracks detection. The basic working principles of neural networks are presented. Experimental vibration signals of rotors with and without a propagating crack were used to train the Multi-layer Feed-forward Neural Networks using back-propagation algorithm. The trained neural networks were tested with other set of vibration data. A simple two-layer feed-forward neural network with two neurons in the input layer and one neuron in the output layer trained with the signals of a cracked rotor and a normal rotor without a crack was found to be satisfactory in detecting a propagating crack. Trained three-layer networks were able to detect both the propagating and non-propagating cracks. The FFT of the vibration signals showing variation in amplitude of the harmonics as time progresses are also presented for comparison.
7
Pryor, Connor, Charles Dickens, Eriq Augustine, Alon Albalak, William Yang Wang, and Lise Getoor. "NeuPSL: Neural Probabilistic Soft Logic." In Thirty-Second International Joint Conference on Artificial Intelligence {IJCAI-23}. California: International Joint Conferences on Artificial Intelligence Organization, 2023. http://dx.doi.org/10.24963/ijcai.2023/461.
In this paper, we introduce Neural Probabilistic Soft Logic (NeuPSL), a novel neuro-symbolic (NeSy) framework that unites state-of-the-art symbolic reasoning with the low-level perception of deep neural networks. To model the boundary between neural and symbolic representations, we propose a family of energy-based models, NeSy Energy-Based Models, and show that they are general enough to include NeuPSL and many other NeSy approaches. Using this framework, we show how to seamlessly integrate neural and symbolic parameter learning and inference in NeuPSL. Through an extensive empirical evaluation, we demonstrate the benefits of using NeSy methods, achieving upwards of 30% improvement over independent neural network models. On a well-established NeSy task, MNIST-Addition, NeuPSL demonstrates its joint reasoning capabilities by outperforming existing NeSy approaches by up to 10% in low-data settings. Furthermore, NeuPSL achieves a 5% boost in performance over state-of-the-art NeSy methods in a canonical citation network task with up to a 40 times speed up.
Neural network models for artificial intelligence offer an approach fundamentally different from conventional symbolic approaches, but the merits of the two paradigms cannot be fairly compared until neural network models with large numbers of ”neurons” are implemented. Despite the attractiveness of neural networks for computing applications which involve adaptation and learning, most of the published demonstrations of neural network technology have involved relatively small numbers of ”neurons”. One reason for this is the poor match between conventional electronic serial or coarse-grained multiple-processor computers and the massive parallelism and communication requirements of neural network models. The self-pumped optical neural network (SPONN) described here is a fine-grained optical architecture which features massive parallelism and a much greater degree of interconnectivity than bus-oriented or hypercube electronic architectures. SPONN is potentially capable of implementing neural networks consisting of 105-106 neurons with 109-1010 interconnections. The mapping of neural network models onto the architecture occurs naturally without the need for multiplexing neurons or dealing with contention, routing, and communication bottleneck problems. This simplifies the programming involved compared to electronic implementations.
9
Spall, James, Xianxin Guo, and A. I. Lvovsky. "Hybrid training of optical neural networks." In Frontiers in Optics. Washington, D.C.: Optica Publishing Group, 2022. http://dx.doi.org/10.1364/fio.2022.ftu6d.2.
Optical neural networks are often trained “in-silico” on digital simulators, but physical imperfections that cannot be modelled may lead to a “reality gap” between the simulator and the physical system. In this work we present hybrid training, where the weight matrix is trained by computing neuron values optically using the actual physical network.
10
Xie, Xuan, Kristian Kersting, and Daniel Neider. "Neuro-Symbolic Verification of Deep Neural Networks." In Thirty-First International Joint Conference on Artificial Intelligence {IJCAI-22}. California: International Joint Conferences on Artificial Intelligence Organization, 2022. http://dx.doi.org/10.24963/ijcai.2022/503.
Formal verification has emerged as a powerful approach to ensure the safety and reliability of deep neural networks. However, current verification tools are limited to only a handful of properties that can be expressed as first-order constraints over the inputs and output of a network. While adversarial robustness and fairness fall under this category, many real-world properties (e.g., "an autonomous vehicle has to stop in front of a stop sign") remain outside the scope of existing verification technology. To mitigate this severe practical restriction, we introduce a novel framework for verifying neural networks, named neuro-symbolic verification. The key idea is to use neural networks as part of the otherwise logical specification, enabling the verification of a wide variety of complex, real-world properties, including the one above. A defining feature of our framework is that it can be implemented on top of existing verification infrastructure for neural networks, making it easily accessible to researchers and practitioners.
Reports on the topic "Neural networks":
1
Smith, Patrick I. Neural Networks. Office of Scientific and Technical Information (OSTI), September 2003. http://dx.doi.org/10.2172/815740.
Johnson, John L., and C. C. Sung. Neural Networks. Fort Belvoir, VA: Defense Technical Information Center, January 1990. http://dx.doi.org/10.21236/ada222110.
Tarasenko, Andrii O., Yuriy V. Yakimov, and Vladimir N. Soloviev. Convolutional neural networks for image classification. [б. в.], February 2020. http://dx.doi.org/10.31812/123456789/3682.
This paper shows the theoretical basis for the creation of convolutional neural networks for image classification and their application in practice. To achieve the goal, the main types of neural networks were considered, starting from the structure of a simple neuron to the convolutional multilayer network necessary for the solution of this problem. It shows the stages of the structure of training data, the training cycle of the network, as well as calculations of errors in recognition at the stage of training and verification. At the end of the work the results of network training, calculation of recognition error and training accuracy are presented.
4
Holder, Nanette S. Introduction to Neural Networks. Fort Belvoir, VA: Defense Technical Information Center, March 1992. http://dx.doi.org/10.21236/ada248258.
Abu-Mostafa, Yaser S., and Amir F. Atiya. Theory of Neural Networks. Fort Belvoir, VA: Defense Technical Information Center, July 1991. http://dx.doi.org/10.21236/ada253187.
Alltop, W. O. Piecewise Linear Neural Networks. Fort Belvoir, VA: Defense Technical Information Center, August 1992. http://dx.doi.org/10.21236/ada265031.
Wiggins, Vince L., Larry T. Looper, and Sheree K. Engquist. Neural Networks: A Primer. Fort Belvoir, VA: Defense Technical Information Center, May 1991. http://dx.doi.org/10.21236/ada235920.
Yu, Haichao, Haoxiang Li, Honghui Shi, Thomas S. Huang, and Gang Hua. Any-Precision Deep Neural Networks. Web of Open Science, December 2020. http://dx.doi.org/10.37686/ejai.v1i1.82.
We present Any-Precision Deep Neural Networks (Any- Precision DNNs), which are trained with a new method that empowers learned DNNs to be flexible in any numerical precision during inference. The same model in runtime can be flexibly and directly set to different bit-width, by trun- cating the least significant bits, to support dynamic speed and accuracy trade-off. When all layers are set to low- bits, we show that the model achieved accuracy compara- ble to dedicated models trained at the same precision. This nice property facilitates flexible deployment of deep learn- ing models in real-world applications, where in practice trade-offs between model accuracy and runtime efficiency are often sought. Previous literature presents solutions to train models at each individual fixed efficiency/accuracy trade-off point. But how to produce a model flexible in runtime precision is largely unexplored. When the demand of efficiency/accuracy trade-off varies from time to time or even dynamically changes in runtime, it is infeasible to re-train models accordingly, and the storage budget may forbid keeping multiple models. Our proposed framework achieves this flexibility without performance degradation. More importantly, we demonstrate that this achievement is agnostic to model architectures. We experimentally validated our method with different deep network backbones (AlexNet-small, Resnet-20, Resnet-50) on different datasets (SVHN, Cifar-10, ImageNet) and observed consistent results.
9
Keller, P. E. Artificial neural networks in medicine. Office of Scientific and Technical Information (OSTI), July 1994. http://dx.doi.org/10.2172/10162484.
Johnson, M. A., G. Kendall, P. J. Cote, and L. V. Meisel. Neural Networks in Seizure Diagnosis. Fort Belvoir, VA: Defense Technical Information Center, March 1995. http://dx.doi.org/10.21236/ada295629.
