The Experts below are selected from a list of 216 Experts worldwide ranked by ideXlab platform
Ivo Bukovsky - One of the best experts on this subject based on the ideXlab platform.
-
Potentials of Quadratic Neural Unit for Applications
Advances in Abstract Intelligence and Soft Computing, 2020Co-Authors: Ricardo Rodriguez, Ivo Bukovsky, Noriyasu HommaAbstract:The paper discusses the quadratic Neural Unit (QNU) and highlights its attractiveness for industrial applications such as for plant modeling, control, and time series prediction. Linear systems are still often preferred in industrial control applications for their solvable and single solution nature and for the clarity to the most application engineers. Artificial Neural networks are powerful cognitive nonlinear tools, but their nonlinear strength is naturally repaid with the local minima problem, overfitting, and high demands for application-correct Neural architecture and optimization technique that often require skilled users. The QNU is the important midpoint between linear systems and highly nonlinear Neural networks because the QNU is relatively very strong in nonlinear approximation; however, its optimization and performance have fast and convex-like nature, and its mathematical structure and the derivation of the learning rules is very comprehensible and efficient for implementation. These advantages of QNU are demonstrated by using real and theoretical examples.
-
HONU and Supervised Learning Algorithms in Adaptive Feedback Control
Advances in Computational Intelligence and Robotics, 2020Co-Authors: Peter Mark Benes, Miroslav Erben, Martin Vesely, Ondrej Liska, Ivo BukovskyAbstract:This chapter is a summarizing study of Higher Order Neural Units featuring the most common learning algorithms for identification and adaptive control of most typical representatives of plants of single-input single-output (SISO) nature in the control engineering field. In particular, the linear Neural Unit (LNU, i.e., 1st order HONU), quadratic Neural Unit (QNU, i.e. 2nd order HONU), and cubic Neural Unit (CNU, i.e. 3rd order HONU) will be shown as adaptive feedback controllers of typical models of linear plants in control including identification and control of plants with input time delays. The investigated and compared learning algorithms for HONU will be the step-by-step Gradient Descent adaptation with the study of known modifications of learning rate for improved convergence, the batch Levenberg-Marquardt algorithm, and the Resilient Back-Propagation algorithm. The theoretical achievements will be summarized and discussed as regards their usability and the real issues of control engineering tasks.
-
IWCIM - Adaptive polynomial filters with individual learning rates for computationally efficient lung tumor motion prediction
2015 International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM), 2015Co-Authors: Matous Cejnek, Ivo Bukovsky, Noriyasu Homma, Ondrej LiskaAbstract:This paper presents a study of higher-order Neural Units as polynomial adaptive filters with multiple-learning-rate gradient descent for 3-D lung tumor motion prediction. The method is compared with single-learning rate gradient descent approaches with and without learning rate normalization. Experimental analysis is done with linear and quadratic Neural Unit. The influence of correct selection of adaptation parameters and the dependence of learning time on accuracy were experimentally analyzed. The prediction accuracy is nearly equal to recently published results of batch retraining approaches while the computational efficiency is higher for the introduced approach.
-
Neural NETWORK APPROACH TO RAILWAY STAND LATERAL SKEW CONTROL
Computer Science & Information Technology ( CS & IT ), 2014Co-Authors: Peter Mark Benes, Ivo Bukovsky, Matous Cejnek, Jan KalivodaAbstract:The paper presents a study of an adaptive approach to lateral skew control for an experimental railway stand. The preliminary experiments with the real experimental railway stand and simulations with its 3-D mechanical model, indicates difficulties of model-based control of the device. Thus, use of Neural networks for identification and control of lateral skew shall be investigated. This paper focuses on real-data based modelling of the railway stand by various Neural network models, i.e; linear Neural Unit and quadratic Neural Unit architectures. Furthermore, training methods of these Neural architectures as such, real-time-recurrentlearning and a variation of back-propagation-through-time are examined, accompanied by a discussion of the produced experimental results.
-
Lung Tumor Motion Prediction by static Neural networks
2012Co-Authors: Richardo Rodriguez, Jiri Bila, Ivo Bukovsky, Noriyasu HommaAbstract:This paper presents a study of lung tumor-motion time-series prediction, first, with the use of conventional static (feedforward) MLP Neural network (with a single hidden perceptron layer) and, second, with the static quadratic Neural Unit (QNU), i.e., a class of polynomial Neural network (or a higher-order Neural Unit). We also demonstrate that QNU can be trained in a very efficient and fast way for real time retraining due to its linear nature of optimization problem. The objective is the prediction accuracy of 1 [mm] for 1-second prediction horizon. So it is well applicable for radiation tracking therapy.
Jiri Bila - One of the best experts on this subject based on the ideXlab platform.
-
3PGCIC - Prediction of Highly Non-stationary Time Series Using Higher-Order Neural Units
Advances on P2P Parallel Grid Cloud and Internet Computing, 2017Co-Authors: Ricardo Rodríguez Jorge, Jiri Bila, Edgar A. Martinez Garcia, Jolanta Mizera-pietraszko, Rafael Torres CórdobaAbstract:Adaptive predictive models can use conventional and nonconventional Neural networks for highly non-stationary time series prediction. However, conventional Neural networks present a series of known drawbacks. This paper presents a brief discussion about this concern as well as how the basis of higher-order Neural Units can overcome some of them; it also describes a sliding window technique alongside the batch optimization technique for capturing the dynamics of non-stationary time series over a Quadratic Neural Unit, a special case of higher-order Neural Units. Finally, an experimental analysis is presented to demonstrate the effectiveness of the proposed approach.
-
FGCT - Arrhythmia disease classification using a higher-order Neural Unit
2015 Fourth International Conference on Future Generation Communication Technology (FGCT), 2015Co-Authors: Ricardo Rodriguez, Osslan Osiris Vergara Villegas, Vianey Guadalupe Cruz Sánchez, Jiri Bila, Adriana MexicanoAbstract:This paper presents a quadratic Neural Unit with error backpropagation learning algorithm to classify electrocardiogram arrhythmia disease. The electrocardiogram arrhythmia classification scheme consists of data acquisition, feature extraction, feature reduction, and a quadratic Neural Unit classifier to discriminate three different types of arrhythmia. A total of 44 records were obtained from MIT-BIH arrhythmia database to test the efficiency of arrhythmia disease classification method, the obtained results were a specificity of 97.60 % and a sensitivity of 97.05 %. The best accuracy classification rate obtained using the presented approach has been of 98.16 %.
-
Lung Tumor Motion Prediction by static Neural networks
2012Co-Authors: Richardo Rodriguez, Jiri Bila, Ivo Bukovsky, Noriyasu HommaAbstract:This paper presents a study of lung tumor-motion time-series prediction, first, with the use of conventional static (feedforward) MLP Neural network (with a single hidden perceptron layer) and, second, with the static quadratic Neural Unit (QNU), i.e., a class of polynomial Neural network (or a higher-order Neural Unit). We also demonstrate that QNU can be trained in a very efficient and fast way for real time retraining due to its linear nature of optimization problem. The objective is the prediction accuracy of 1 [mm] for 1-second prediction horizon. So it is well applicable for radiation tracking therapy.
-
quadratic Neural Unit and its network in validation of process data of steam turbine loop and energetic boiler
International Joint Conference on Neural Network, 2010Co-Authors: Ivo Bukovsky, Martin Lepold, Jiri BilaAbstract:This paper discusses results and advantages of the application of quadratic Neural Units and novel quadratic Neural network to modeling of real data for purposes of validation of measured data in energetic processes. A feed forward network of quadratic Neural Units (a class of higher order Neural network) with sequential learning is presented. This quadratic network with this learning technique reduces computational time for models with large number of inputs, sustains optimization convexity of a quadratic model, and also displays sufficient nonlinear approximation capability for the real processes. A comparison of performances of the quadratic Neural Units, quadratic Neural networks, and the use of common multilayer feed forward Neural networks all trained by Levenberg-Marquard algorithm is discussed.
-
IJCNN - Quadratic Neural Unit and its network in validation of process data of steam turbine loop and energetic boiler
The 2010 International Joint Conference on Neural Networks (IJCNN), 2010Co-Authors: Ivo Bukovsky, Martin Lepold, Jiri BilaAbstract:This paper discusses results and advantages of the application of quadratic Neural Units and novel quadratic Neural network to modeling of real data for purposes of validation of measured data in energetic processes. A feed forward network of quadratic Neural Units (a class of higher order Neural network) with sequential learning is presented. This quadratic network with this learning technique reduces computational time for models with large number of inputs, sustains optimization convexity of a quadratic model, and also displays sufficient nonlinear approximation capability for the real processes. A comparison of performances of the quadratic Neural Units, quadratic Neural networks, and the use of common multilayer feed forward Neural networks all trained by Levenberg-Marquard algorithm is discussed.
Madan M. Gupta - One of the best experts on this subject based on the ideXlab platform.
-
DevelopmentofHigher-OrderNeuralUnits forControland PatternRecognition*
2020Co-Authors: Madan M. GuptaAbstract:The computational Neural-network structures described intheliterature areoften based onthenotion of linear Neural Units (LNUs). Thebiological neurons consist ofcomplex computing elements, whichperform more computations than justlinearsummation.The computational efficiency oftheNeural networks depends on their structure andthetraining methods employed. Higher- order combinations ofinputs andweights will yield higher Neural performance. Inthis paper, aquadratic-Neural Unit (QNU)hasbeendeveloped using a novel general matrix formofthequadratic operation. WehaveusedtheQNUfor realizing different logic circuits. IndexTerms-HigherorderNeuralUnits (HONU),Neural networks, Pattern classification, Quadratic function.
-
Computational Structure of Correlative Type Higher Order Neural Units with Applications
2008 IEEE Region 10 and the Third international Conference on Industrial and Information Systems, 2008Co-Authors: Madan M. GuptaAbstract:The computational Neural-network structures described in the literature are often based on the concept of linear synaptic operations. In biological processes, however, neurons form a set of very complex computing elements and perform much more complex computations than just the linear aggregation. It is well known that the computational efficiency of Neural networks depends on its morphology and their learning and adaptation strategies. In our engineering design and economic processes, however, Neural inputs are not necessarily the independent quantities rather they form some correlative attributes. In this paper we present a new class of correlative type higher-order Neural Units (HONUs) with nonlinear combinations of inputs and weights. In particular, in this paper we present a quadratic-Neural Unit (QNU) and a cubic-Neural Unit (CNU), and their higher-order extensions. For illustrating the applications of these correlative type higher-order Neural Units, we have given some examples taken from the field of feedback control systems and logic circuits. It has been noticed that these higher-order correlative types of Neural Units, both static and dynamic, can be used for many applications in various fields such as engineering, medical imaging and economics.
-
Neural Units with Higher-Order Synaptic Operations for Robotic Image Processing Applications
Soft Computing, 2006Co-Authors: Ki-young Song, Madan M. GuptaAbstract:Neural Units with higher-order synaptic operations have good computational properties in information processing and control applications. This paper presents Neural Units with higher-order synaptic operations for visual image processing applications. We use the Neural Units with higher-order synaptic operations for edge detection and employ the Hough transform to process the edge detection results. The edge detection method based on the Neural Unit with higher-order synaptic operations has been applied to solve routing problems of mobile robots. Simulation results show that the proposed Neural Units with higher-order synaptic operations are efficient for image processing and routing applications of mobile robots.
-
Development of higher-order Neural Units for control and pattern recognition
NAFIPS 2005 - 2005 Annual Meeting of the North American Fuzzy Information Processing Society, 2005Co-Authors: Madan M. GuptaAbstract:The computational Neural-network structures described in the literature are often based on the notion of linear Neural Units (LNUs). The biological neurons consist of complex computing elements, which perform more computations than just linear summation. The computational efficiency of the Neural networks depends on their structure and the training methods employed. Higher-order combinations of inputs and weights will yield higher Neural performance. In this paper, a quadratic-Neural Unit (QNU) has been developed using a novel general matrix form of the quadratic operation. We have used the QNU for realizing different logic circuits.
-
Cubic Neural Unit for control applications
Fourth International Symposium on Uncertainty Modeling and Analysis 2003. ISUMA 2003., 2003Co-Authors: Ki-young Song, S. Redlapalli, Madan M. GuptaAbstract:A novel Neural structure called a cubic Neural Unit (CNU) is developed. The CNU can be used as a nonlinear controller for controlling the continuous dynamic plants. The structure and the mathematical details of the CNU are described briefly, and an example is used to show the usefulness of the CNU as a nonlinear controller. Neural structures, linear and quadratic, are used as state controllers for complex control systems such as satellite control. Results of all Neural structures as neurocontrollers are shown
James J Dicarlo - One of the best experts on this subject based on the ideXlab platform.
-
performance optimized hierarchical models predict Neural responses in higher visual cortex
Proceedings of the National Academy of Sciences of the United States of America, 2014Co-Authors: Daniel Yamins, Ha Hong, Charles F Cadieu, Ethan A Solomon, Darren Seibert, James J DicarloAbstract:The ventral visual stream underlies key human visual object recognition abilities. However, Neural encoding in the higher areas of the ventral stream remains poorly understood. Here, we describe a modeling approach that yields a quantitatively accurate model of inferior temporal (IT) cortex, the highest ventral cortical area. Using high-throughput computational techniques, we discovered that, within a class of biologically plausible hierarchical Neural network models, there is a strong correlation between a model’s categorization performance and its ability to predict individual IT Neural Unit response data. To pursue this idea, we then identified a high-performing Neural network that matches human performance on a range of recognition tasks. Critically, even though we did not constrain this model to match Neural data, its top output layer turns out to be highly predictive of IT spiking responses to complex naturalistic images at both the single site and population levels. Moreover, the model’s intermediate layers are highly predictive of Neural responses in the V4 cortex, a midlevel visual area that provides the dominant cortical input to IT. These results show that performance optimization—applied in a biologically appropriate model class—can be used to build quantitative predictive models of Neural processing.
Sean E. Shaheen - One of the best experts on this subject based on the ideXlab platform.
-
Neurons in Polymer: Hardware Neural Units Based on Polymer Memristive Devices and Polymer Transistors
IEEE Transactions on Electron Devices, 2014Co-Authors: Robert A. Nawrocki, Richard M. Voyles, Sean E. ShaheenAbstract:We present here incremental steps toward realizing a tangible polymer neuromorphic architecture in the form of McCulloch-Pitts (nonspiking) neurons made from polymer electronics components, namely, memristive read-only-memory devices, transistors, and resistors. In the implementation, the polymer memristive devices perform the equivalent of synaptic weighting, while a polymer resistor subcircuit performs the equivalent of somatic summing. The sum is sent to a single transistor to apply the activation function. The complete circuit approximates the function of a single Neural Unit, which would form the basis for a hardware artificial Neural network. It is shown here that a single, two-input Unit, fit with three memristive devices per input, can perform continuous value classification applied to an active tether application, with a maximum error of 5%.
