Visible to the public Neural Networks

SoS Newsletter- Advanced Book Block

Neural Networks

Artificial neural networks have been used to solve a wide variety of tasks that are hard to solve using ordinary rule-based programming. What has attracted much interest in neural networks is the possibility of learning. Tasks such as function approximation, classification pattern and sequence recognition, anomoly detection, filtering, clustering, blind source separation and compression and controls all have security implications. The work cited here looks at authentication, use of learning to develop a hybrid security system, and attack behavior classification in artificial neural networks.

  • Baheti, Ankita; Singh, Lokesh; Khan, Asif Ullah, "Proposed Method for Multimedia Data Security Using Cyclic Elliptic Curve, Chaotic System, and Authentication Using Neural Network," Communication Systems and Network Technologies (CSNT), 2014 Fourth International Conference on , vol., no., pp.664,668, 7-9 April 2014. (ID#:14-1710) Available at: As multimedia applications are used increasingly, security becomes an important issue of security of images. The combination of chaotic theory and cryptography forms an important field of information security. In the past decade, chaos based image encryption is given much attention in the research of information security and a lot of image encryption algorithms based on chaotic maps have been proposed. But, most of them delay the system performance, security, and suffer from the small key space problem. This paper introduces an efficient symmetric encryption scheme based on a cyclic elliptic curve and chaotic system that can overcome these disadvantages. The cipher encrypts 256-bit of plain image to 256-bit of cipher image within eight 32-bit registers. The scheme generates pseudorandom bit sequences for round keys based on a piecewise nonlinear chaotic map. Then, the generated sequences are mixed with the key sequences derived from the cyclic elliptic curve points. The proposed algorithm has good encryption effect, large key space, high sensitivity to small change in secret keys and fast compared to other competitive algorithms. Keywords: Authentication; Chaotic communication; Elliptic curves; Encryption; Media; Multimedia communication; authentication; chaos; decryption; encryption; neural network
  • Singh, Nikita; Chandra, Nidhi, "Integrating Machine Learning Techniques to Constitute a Hybrid Security System," Communication Systems and Network Technologies (CSNT), 2014 Fourth International Conference on , vol., no., pp.1082,1087, 7-9 April 2014. (ID#:14-1711) Available at: Computer Security has been discussed and improvised in many forms and using different techniques as well as technologies. The enhancements keep on adding as the security remains the fastest updating unit in a computer system. In this paper we propose a model for securing the system along with the network and enhance it more by applying machine learning techniques SVM (support vector machine) and ANN (Artificial Neural Network). Both the techniques are used together to generate results which are appropriate for analysis purpose and thus, prove to be the milestone for security. Keywords: Artificial neural networks; Intrusion detection; Neurons; Probabilistic logic; Support vector machines; Training; Artificial neural network; Host logs; Machine Learning; Network logs; Support vector machine
  • Abdul Razzaq, Khalid Latif, H. Farooq Ahmad, Ali Hur, Zahid Anwar, Peter Charles Bloodsworth, "Semantic Security Against Web Application Attacks," Information Sciences: an International Journal, Volume 254, January, 2014. (ID#:14-1712) Available at: This paper proposes an ontology-based method for the detection and identification of web application attacks, including zero day attacks, with few false positives. This ontology-based solution, as opposed to current signature-based methods, classifies web application attacks by employing semantic rules to identify the application context, probable attacks, and protocol used. The rules allow detection of complex variations of web application attacks, as well as provide for a platform and technology independent system. Keywords: Application security, Semantic rule engine, Semantic security
  • Al-Jarrah, Omar; Arafat, Ahmad, "Network Intrusion Detection System Using Attack Behavior Classification," Information and Communication Systems (ICICS), 2014 5th International Conference on , vol., no., pp.1,6, 1-3 April 2014. (ID#:14-1713) Available at: Intrusion Detection Systems (IDS) have become a necessity in computer security systems because of the increase in unauthorized accesses and attacks. Intrusion Detection is a major component in computer security systems that can be classified as Host-based Intrusion Detection System (HIDS), which protects a certain host or system and Network-based Intrusion detection system (NIDS), which protects a network of hosts and systems. This paper addresses Probes attacks or reconnaissance attacks, which try to collect any possible relevant information in the network. Network probe attacks have two types: Host Sweep and Port Scan attacks. Host Sweep attacks determine the hosts that exist in the network, while port scan attacks determine the available services that exist in the network. This paper uses an intelligent system to maximize the recognition rate of network attacks by embedding the temporal behavior of the attacks into a TDNN neural network structure. The proposed system consists of five modules: packet capture engine, preprocessor, pattern recognition, classification, and monitoring and alert module. We have tested the system in a real environment where it has shown good capability in detecting attacks. In addition, the system has been tested using DARPA 1998 dataset with 100% recognition rate. In fact, our system can recognize attacks in a constant time. Keywords: IP networks ;Intrusion detection; Neural networks ;Pattern recognition; Ports (Computers); Probes; Protocols; Host sweep; Intrusion Detection Systems; Network probe attack; Port scan; TDNN neural network
  • Singla, P.; Sachdeva, P.; Ahmad, M., "A Chaotic Neural Network Based Cryptographic Pseudo-Random Sequence Design," Advanced Computing & Communication Technologies (ACCT), 2014 Fourth International Conference on , vol., no., pp.301,306, 8-9 Feb. 2014. (ID#:14-1714) Available at: Efficient random sequence generators are significant in the application areas of cryptographic stream cipher design, statistical sampling and simulation, direct spread spectrum, etc. A cryptographically efficient pseudo-random sequence should have the characteristics of high randomness and encryption effect. The statistical quality of pseudo-random sequences determines the strength of cryptographic system. The generation of pseudo-random sequences with high randomness and encryption effect is a key challenge. A sequence with poor randomness threatens the security of cryptographic system. In this paper, the features and strengths of chaos and neural network are combined to design a novel pseudo-random binary sequence generator for cryptographic applications. The statistical performance of the proposed chaotic neural network based pseudo random sequence generator is examined against the NIST SP800-22 randomness tests and multimedia image encryption. The results of investigations are promising and depict its relevance for cryptographic applications. Keywords: chaos; cryptography; neural nets; random sequences; sampling methods; chaotic neural network based cryptographic pseudo-random sequence design; cryptographic stream cipher design; cryptographic system security; cryptographic system strength determination; direct spread spectrum; encryption effect characteristics; high randomness characteristics; statistical sampling; statistical simulation; Biological neural networks; Chaotic communication; Encryption; Generators; Chaotic; cryptography; image encryption; neural network; pseudo-random sequence generator
  • Khatri, P., "Using identity and trust with key management for achieving security in Ad hoc Networks," Advance Computing Conference (IACC), 2014 IEEE International , vol., no., pp.271,275, 21-22 Feb. 2014. (ID#:14-1715) Available at: Communication in Mobile Ad hoc network is done over a shared wireless channel with no Central Authority (CA) to monitor. Responsibility of maintaining the integrity and secrecy of data, nodes in the network are held responsible. To attain the goal of trusted communication in MANET (Mobile Ad hoc Network) lot of approaches using key management has been implemented. This work proposes a composite identity and trust based model (CIDT) which depends on public key, physical identity, and trust of a node which helps in secure data transfer over wireless channels. CIDT is a modified DSR routing protocol for achieving security. Trust Factor of a node along with its key pair and identity is used to authenticate a node in the network. Experience based trust factor (TF) of a node is used to decide the authenticity of a node. A valid certificate is generated for authentic node to carry out the communication in the network. Proposed method works well for self certification scheme of a node in the network. Keywords: data communication; mobile ad hoc networks routing protocols; telecommunication security; wireless channels; MANET; ad hoc networks; central authority; data integrity; data secrecy; experience based trust factor; identity model; key management; mobile ad hoc network ;modified DSR routing protocol; physical identity; public key; secure data transfer; security; self certification scheme; shared wireless channel; trust factor; trust model; trusted communication; wireless channels; Artificial neural networks; Mobile ad hoc networks; Protocols; Public key; Servers; Certificate; MANET; Public key; Secret key; Trust Model
  • Kumar, D.; Gupta, S.; Sehgal, P., "Comparing gradient based learning methods for optimizing predictive neural networks," Engineering and Computational Sciences (RAECS), 2014 Recent Advances in , vol., no., pp.1,6, 6-8 March 2014a. (ID#:14-1716) Available at: In this paper, we compare the performance of various gradient based techniques in optimizing the neural networks employed for prediction modeling. Training of neural network based predictive models is done using gradient based techniques, which involves searching for the point of minima on multidimensional energy function by providing step-wise corrective adjustment of weight vector present in hidden layers. Convergence of different gradient techniques is studied and compared by performing experiments in neural network toolbox package of MATLAB. Bulky data sets extracted from live data warehouse of life insurance sector are employed with gradient methods for developing the predictive models. Convergence behaviors of learning methods - gradient descent method, Levenberg Marquardt method, conjugate gradient method and scaled conjugate gradient method have been observed. Keywords: conjugate gradient methods; data warehouses; learning (artificial intelligence); neural nets; Levenberg Marquardt method; MATLAB; bulky data sets; convergence behavior; gradient based learning method; gradient based techniques; gradient descent method; gradient techniques; life insurance sector; live data warehouse; multidimensional energy function; neural network based predictive models; neural network toolbox package; prediction modeling; predictive neural networks; scaled conjugate gradient method; step-wise corrective adjustment; weight vector; Convergence; Gradient methods; Neural networks; Neurons; Predictive models; Training; Vectors; conjugate gradient; gradient methods; learning algorithms; neural networks; nonlinear optimization
  • Zhang, H.; Wang, Z.; Liu, D., "A Comprehensive Review of Stability Analysis of Continuous-Time Recurrent Neural Networks," Neural Networks and Learning Systems, IEEE Transactions on , vol.25, no.7, pp.1229,1262, July 2014. (ID#:14-1717) Available at: Stability problems of continuous-time recurrent neural networks have been extensively studied, and many papers have been published in the literature. The purpose of this paper is to provide a comprehensive review of the research on stability of continuous-time recurrent neural networks, including Hopfield neural networks, Cohen-Grossberg neural networks, and related models. Since time delay is inevitable in practice, stability results of recurrent neural networks with different classes of time delays are reviewed in detail. For the case of delay-dependent stability, the results on how to deal with the constant/variable delay in recurrent neural networks are summarized. The relationship among stability results in different forms, such as algebraic inequality forms, (M) -matrix forms, linear matrix inequality forms, and Lyapunov diagonal stability forms, is discussed and compared. Some necessary and sufficient stability conditions for recurrent neural networks without time delays are also discussed. Concluding remarks and future directions of stability analysis of recurrent neural networks are given. Keywords: Biological neural networks; Delays; Neurons; ecurrent neural networks; Stability criteria;(M) -matrix; Cohen--Grossberg neural networks; Cohen-Grossberg neural networks; Hopfield neural networks; Lyapunov diagonal stability (LDS); M-matrix; discrete delay; distributed delays; linear matrix inequality (LMI); recurrent neural networks; robust stability; stability
  • Yu Wang; Boxun Li; Rong Luo; Yiran Chen; Ningyi Xu; Huazhong Yang, "Energy efficient neural networks for big data analytics," Design, Automation and Test in Europe Conference and Exhibition (DATE), 2014 , vol., no., pp.1,2, 24-28 March 2014. (ID#:14-1718) Available at: The world is experiencing a data revolution to discover knowledge in big data. Large scale neural networks are one of the mainstream tools of big data analytics. Processing big data with large scale neural networks includes two phases: the training phase and the operation phase. Huge computing power is required to support the training phase. And the energy efficiency (power efficiency) is one of the major considerations of the operation phase. We first explore the computing power of GPUs for big data analytics and demonstrate an efficient GPU implementation of the training phase of large scale recurrent neural networks (RNNs). We then introduce a promising ultrahigh energy efficient implementation of neural networks' operation phase by taking advantage of the emerging memristor technique. Experiment results show that the proposed GPU implementation of RNNs is able to achieve 2 ~ 11x speed-up compared with the basic CPU implementation. And the scaled-up recurrent neural network trained with GPUs realizes an accuracy of 47% on the Microsoft Research Sentence Completion Challenge, the best result achieved by a single RNN on the same dataset. In addition, the proposed memristor-based implementation of neural networks demonstrates power efficiency of > 400 GFLOPS/W and achieves energy savings of 22x on the HMAX model compared with its pure digital implementation counterpart. Keywords: data analysis; electronic engineering computing; graphics processing units; memristors; recurrent neural nets; CPU implementation; GPU implementation; HMAX model; RNNs; big data analytics; energy efficient neural networks; large scale recurrent neural networks; memristor technique; neural networks operation phase; neural networks training phase; power efficiency; Data handling; Data storage systems; Information management; Memristors; Recurrent neural networks; Training
  • Rakkiyappan, R.; Cao, J.; Velmurugan, G., "Existence and Uniform Stability Analysis of Fractional-Order Complex-Valued Neural Networks With Time Delays," Neural Networks and Learning Systems, IEEE Transactions on, vol. PP, no.99, pp.1,1, March 2014. (ID#:14-1719) Available at: This paper deals with the problem of existence and uniform stability analysis of fractional-order complex-valued neural networks with constant time delays. Complex-valued recurrent neural networks is an extension of real-valued recurrent neural networks that includes complex-valued states, connection weights, or activation functions. This paper explains sufficient condition for the existence and uniform stability analysis of such networks. Three numerical simulations are delineated to substantiate the effectiveness of the theoretical results. Keywords: Artificial neural networks; Biological neural networks; Delay effects; Mathematics; Recurrent neural networks; Stability analysis; Banach contraction fixed point theorem; complex-valued neural networks; fractional order; time delays
  • Alfaro-Ponce, M.; Arguelles Cruz, A.; Chairez, I., "Adaptive Identifier for Uncertain Complex Nonlinear Systems Based on Continuous Neural Networks," Neural Networks and Learning Systems, IEEE Transactions on , vol.25, no.3, pp.483,494, March 2014. (ID#:14-1720) Available at: This paper presents the design of a complex-valued differential neural network identifier for uncertain nonlinear systems defined in the complex domain. This design includes the construction of an adaptive algorithm to adjust the parameters included in the identifier. The algorithm is obtained based on a special class of controlled Lyapunov functions. The quality of the identification process is characterized using the practical stability framework. Indeed, the region where the identification error converges is derived by the same Lyapunov method. This zone is defined by the power of uncertainties and perturbations affecting the complex-valued uncertain dynamics. Moreover, this convergence zone is reduced to its lowest possible value using ideas related to the so-called ellipsoid methodology. Two simple but informative numerical examples are developed to show how the identifier proposed in this paper can be used to approximate uncertain nonlinear systems valued in the complex domain. Keywords: Lyapunov methods; identification; neural nets; nonlinear systems; uncertain systems; Lyapunov method; adaptive algorithm; adaptive identifier; approximate uncertain nonlinear systems; complex domain; complex valued differential neural network identifier; complex-valued uncertain dynamics; continuous neural networks; controlled Lyapunov functions; convergence zone; ellipsoid methodology; identification error; identification process; practical stability framework; uncertain complex nonlinear systems; Artificial neural networks; Biological neural networks; Least squares approximations; Lyapunov methods; Nonlinear systems; Training; Complex-valued neural networks; continuous neural network; controlled Lyapunov function; nonparametric identifier


Articles listed on these pages have been found on publicly available internet pages and are cited with links to those pages. Some of the information included herein has been reprinted with permission from the authors or data repositories. Direct any requests via Email to SoS.Project (at) for removal of the links or modifications to specific citations. Please include the ID# of the specific citation in your correspondence.