Hybrid neural network

About: Hybrid neural network is a research topic. Over the lifetime, 1305 publications have been published within this topic receiving 18223 citations.

Exponential Stabilization for Hybrid Recurrent Neural Networks by Delayed Noises Rooted in Discrete Observations of State and Mode

Abstract: Recently, the random noises derived from discrete state observations are creatively designed to realize the role of stabilization for deterministic systems in the existing result. However, for a hybrid neural network, except for the factor of discrete state observations, one always needs to consider the factors of delays and discrete mode identifications. Hence, taking delays and discrete mode identifications into account for random noises is more reasonable and practical than the original work. Motivated by the idea above, this brief is to design delayed random noises derived from discrete state observations and discrete mode identifications to almost surely exponentially stabilize an unstable hybrid recurrent neural networks, by virtue of M-matrix and stochastic analysis methods.

Design and implementation of a parallel software for hybrid neural network computation in PVM environment

Abstract: This paper discusses the design of a hybrid artificial neural network (ANN) system and its implementation in the Parallel Virtual Machine (PVM) environment. First, the PVM functions for supporting parallel applications and communications among multiple processes and multiple machines are investigated. Then, the design and construction of a hybrid ANN simulation software is proposed. It includes the user interface, control and SPMD computing levels. The software can be used for supporting parallel simulation of different kinds of learning algorithms and neural computing models.

Technique of Metals Strength Properties Diagnostics Based on the Complex Use of Fuzzy Inference System and Hybrid Neural Network

Abstract: The results of the research concerning development of the technique of metals strength properties diagnostics using combination of the methods of non-destructive control based on the complex use of fuzzy inference system and hybrid neural network are presented in the paper. The acoustic non-destructive control method, the electro-magnetic method and hardness control were used as the control methods within the framework of the proposed technique. The selection of the optimal combination of the methods was performed using fuzzy inference system, in which, the final solution was taken applying Harrington desirability function. The metal strength properties were determined using hybrid neural network the basis of which are fuzzy neurons. The simulation results with the use of samples of Y8 steel have shown that the combination of acoustic and electromegnetic methods of non-destructive testing is an optimal in terms of maximum value of heneral Harrington desiribility index and the hybrid neural network with two layers of neurons and triangular membership functions with combine algorithm of network training is an optimal one in terms of relative error of metals strength properties evaluation. To our mind, the proposed technique may allow us to increase the exactness of metals strength properties determination when the non-destructive methods of control are applied.

Novel Network Intrusion Detection System using Hybrid Neural Network (Hopfield and Kohonen SOM with Conscience Function)

Abstract: Summary Intrusion detection technology is an effective approach to dealing with the problems of network security. In this paper, it presents an intrusion detection model based on hybrid neural network and SVM. The key idea is to aim at taking advantage of classification abilities of neural network for unknown attacks and the expertbased system for the known attacks. We employ data from the third international knowledge discovery and data mining tools competition (KDDcup’99) to train and test the feasibility of our proposed neural network component. According to the results of our experiment, our model achieves 97.2 percent detection rate for DOS and Probing intrusions, and less than 0.04 percent false alarm rate. Expert system can detect R2L and U2R intrusions more accurately than neural network. Therefore, Hybrid model will improve the performance to detect intrusions.

Small area, low power, mixed-mode circuits for hybrid neural network applications

Abstract: 166 Table 1.1 Table 3.1 Table 6.1 VIII