Loading ...
IJTSRD
Education
Volume 4 Issue 5
12
0
Try Now
Log In
Pricing
International Journal of Trend in Scientific Research and Development (IJTSRD) Volume 5 Issue 4, May-June 2021 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470 @ IJTSRD | Unique Paper ID – IJTSRD41171 | Volume – 5 | Issue – 4 | May-June 2021 Page 23 Neural Network & Applications- An Overview Dr. Mukesh Kumar Lalji1; Dr. Ashish Dongre2; Rajeev Gupta3; Dr. Deepak Paliwal4, O. P. Shrivastava5; N. K. Jain6; Mukesh Katariya7; Manoj Sonkusre8; Gaurav Lalji9; Divya Lalji10; Arvind Jain11; AR. Sandhya Ekbote12 1Vice-Principal; 2Principal; 3I/C Acad; 4HOD Mech; 5I/C HOD – Elect; 6I/C Civil; 7HOD; Prod; 8HOD CS; 9,10Alumni CS; 11Mech; 12I/C HOD - Architecture Department of Technical Education, Skill Development & Employment, M.P. Govt., S.V. Polytechnic College, Shyamla Hills, Bhopal, Madhya Pradesh, India ABSTRACT Models of the brain and nervous system for Highly parallel to the Process information much more like the brain than a serial computer like Learning and simple principles with complex behaviour as well Applications are powerful problem solvers and also used in biological models. KEYWORDS: Human brain, Behaviours, Nerv cells, Neuron Introduction: The human brain contains about 10 billion nerve cells (neurons), Each neuron is connected to the others through 10000 synapses and Properties of the brain. It can learn, reorganize itself from experience and it adapts to the environment. It is robust and fault tolerant, A mathematical model to solve engineering problems. Group of highly connected neurons to realize compositions of non linear functions as in Tasks, Classification, Discrimination/Recognition, Prediction/Estimation with 2 types of networks like Feed forward Neural Networks and Recurrent Neural Networks. Neural Networks: Feed Forward Neural Networks- The information is propagated from the inputs to the outputs and Computations of No non linear functions from n input variables by compositions of Nc algebraic functions with Time has no role (NO cycle between outputs and inputs). How to cite this paper: Dr. Mukesh Kumar Lalji | Dr. Ashish Dongre | Rajeev Gupta | Dr. Deepak Paliwal | O. P. Shrivastava | N. K. Jain | Mukesh Katariya | Manoj Sonkusre | Gaurav Lalji | Divya Lalji | Arvind Jain | AR. Sandhya Ekbote "Neural Network & Applications- An Overview" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456- 6470, Volume-5 | Issue-4, June 2021, pp.23-24, URL: www.ijtsrd.com/papers/ijtsrd41171.pdf Copyright © 2021 by author (s) and International Journal of Trend in Scientific Research and Development Journal. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0) (http: //creativecommons.org/licenses/by/4.0) Recurrent Neural Networks- Can have arbitrary topologies and model systems with internal states (dynamic ones). Delays are associated to a specific weight and Training is more difficult. Performance may be problematic for Stable Outputs may be more difficult to evaluate and unexpected behavior (oscillation, chaos etc). Classification: Class objects in defined categories with Rough decision OR Estimation of the probability for a certain object to belong to a specific class in used Data mining and Applications such as Economy, speech and patterns recognition, sociology, etc. IJTSRD41171 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470 @ IJTSRD | Unique Paper ID – IJTSRD41171 | Volume – 5 | Issue – 4 | May-June 2021 Page 24 Classical neural architectures: Perceptron with Multi-Layer Perceptron and Radial Basis Function (RBF) Kohonen Features maps used in other other architectures such as Neural Networks (Applications). Face recognition, Time series prediction, Process identification, Process control, Optical character recognition and Adaptive filtering. Supervised learning and task The desired response of the neural network in function of particular inputs is well known. One may provide examples and teach the neural network how to fulfill a certain task. Unsupervised learning Idea for group typical input data on the basis of similarity and data clustering. No need of a professor for the network finds itself the correlations between the data for examples of such networks like Kohonen feature maps. Conclusions: Supervised networks are universal approximators (Non recurrent networks) and any limited function can be approximated by a neural network with a finite number of hidden neurons to an arbitrary precision. These are Linear approximators: for a given precision, the number of parameters grows exponentially with the number of variables (polynomials) and Non-linear approximators, the number of parameters grows linearly with the number of variables. Neural networks are utilized as statistical tools such as Adjust non linear functions to fulfill a task and Need of multiple and representative examples but fewer than in other methods with Neural networks enable to model complex static phenomena as well as dynamic ones. Representations of data have to be formulated as Training vectors must be statistically representative of the entire input space with Unsupervised techniques can help the use of The use of Neural Networks needs a good comprehension of the problem. REFERENCE [1] L. Zhang, L. Cheng, H. Li, et al., Hierarchical deep- learning neural networks: Finite elements and beyond, Computational Mechanics 66 (2021) 207– 230. [2] P. Perdikaris, S. Tang, Mechanistic machine learning: Theory, methods, and applications, Theoretical and Applied Mechanics Letters 10 (2020) 141–142. [3] W.K. Liu, G. Karniadakis, S. Tang, et al., A computational mechanics special issue on: Data- driven modeling and simulation-theory, methods, and applications, Computational Mechanics 64 (2019) 275–277. [4] K. Hornik, M. Stinchcombe, H. White, Multilayer feedforward networks are universal approximators, Neural Networks 2 (1989) 359–366. [5] G. Cybenko, Approximation by superpositions of a sigmoidal function, Mathematics of Control, Signal and Systems 2 (1989) 303–314.
