This paper presents a fuzzy hybrid learning algorithm (FHLA) for the radial basis function neural network (RBFNN). The method determines the number of hidden neurons in the RBFNN structure by using cluster validity in...
详细信息
This paper presents a fuzzy hybrid learning algorithm (FHLA) for the radial basis function neural network (RBFNN). The method determines the number of hidden neurons in the RBFNN structure by using cluster validity indices with majority rule while the characteristics of the hidden neurons are initialized based on advanced fuzzy clustering. The FHLA combines the gradient method and the linear least-squared method for adjusting the RBF parameters and the neural network connection weights. The RBFNN with the proposed FHLA is used as a classifier in a face recognition system. The inputs to the RBFNN are the feature vectors obtained by combining shape information and principal component analysis. The designed RBFNN with the proposed FHLA, while providing a faster convergence in the training phase, requires a hidden layer with fewer neurons and less sensitivity to the training and testing patterns. The efficiency of the proposed method is demonstrated on the ORL and Yale face databases, and comparison with other algorithms indicates that the FHLA yields excellent recognition rate in human face recognition. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.
This study considers a joint channel and power allocation for multiple users in underwater acoustic communication networks as a formulated multiplayer MAB game. This study also proposes hierarchical learning algorithm...
详细信息
This study considers a joint channel and power allocation for multiple users in underwater acoustic communication networks as a formulated multiplayer MAB game. This study also proposes hierarchical learning algorithms, which do not need any prior environmental information and direct information exchange among users, to improve the learning ability. In upper sub-learning, each user generates a strategy through the traditional UCB1 strategy. In lower sub-learning, the concept of virtual learning information, which can be obtained as the reward of the last actual played strategy, is introduced to enrich the learning information. Users can enhance their learning ability by learning the outdated virtual learning information in lower sub-learning. As a result, the learning time it takes to achieve the NE is effectively decreased, and the cost of the algorithm is reduced. A distributed optimal NE selection mechanism is proposed to avoid falling into an inadequate local extreme value. Simulation results show high convergence speed and achieved utility of the proposed algorithm. (C) 2018 Elsevier B.V. All rights reserved.
We propose an algorithm for the synthesis of supervisors in discrete event systems. The algorithm is based on a learning algorithm of regular languages proposed by Angluin, and constructs a supervisor realizing an unk...
详细信息
We propose an algorithm for the synthesis of supervisors in discrete event systems. The algorithm is based on a learning algorithm of regular languages proposed by Angluin, and constructs a supervisor realizing an unknown specification which is identified through the interaction between the designer and the algorithm. We also consider the synthesis problem for systems consisting of several processes which behave concurrently. One of serious problems in dealing with such a concurrent system is that the number of states required for describing the global behavior often grows exponentially in the size of the model. To improve this situation, we introduce the concept of dependency defined on the set of events. It prevents the algorithm from considering all interleavings of independent occurrences of events.
In this study, the main objective is to predict the energy needs of residential buildings in the climate zone of Agadir, Morocco, benefitting from orientation, relative compactness, glazing rate, wall surface area, th...
详细信息
In this study, the main objective is to predict the energy needs of residential buildings in the climate zone of Agadir, Morocco, benefitting from orientation, relative compactness, glazing rate, wall surface area, the height and the surface area of the building by using artificial neural networks (ANN) as a learning algorithm. The training data of the neural network were produced using parametric analysis giving rise to 5625 samples in accordance with the mode of construction and use of residential buildings. For each building, it is assumed that the angles of orientation of the samples vary from 0 degrees to 180 degrees, the glazing rates were chosen between 5% and 45%, the heights between 3.5 and 17.5 m and with 25 possible building areas. Three residential buildings "Economic Villa, Economical Construction and Medium Class building" were selected as test data for the neural network model. The Design Builder tool was used for energy demand calculations and a computer program written in Python is used for predictions. As a conclusion;When comparing the calculated values with the outputs of the network, it is proved that the ANN gives satisfactory results with an accuracy of 98.7% and 97.6% for the prediction and test data respectively.
1H-Benzo[ b] pyrrole samples were irradiated in the air with gamma source at 0.969 kGy per hour at room temperature for 24, 48 and 72 h. After irradiation, electron spin resonance, thermogravimetry analysis (TGA) and ...
详细信息
1H-Benzo[ b] pyrrole samples were irradiated in the air with gamma source at 0.969 kGy per hour at room temperature for 24, 48 and 72 h. After irradiation, electron spin resonance, thermogravimetry analysis (TGA) and differential thermal analysis (DTA) measurements were immediately carried out on the irradiated and unirradiated samples. The ESR measurements were performed between 320 and 400 K. ESR spectra were recorded from the samples irradiated for 48 and 72 h. The obtained spectra were observed to be dependent on temperature. Two radical-type centres were detected on the sample. Detected radiation-induced radicals were attributed to R-+center dot NH and R=(center dot)CC2H2. The g-values and hyperfine constants were calculated by means of the experimental spectra. It was also determined from TGA spectrum that both the unirradiated and irradiated samples were decomposed at one step with the rising temperature. Moreover, a theoretical study was presented. Success of the machine learning methods was tested. It was found that bagging techniques, which are widely used in the machine learning literature, could optimise prediction accuracy noticeably.
Context: Recently data-driven program analysis has emerged as a promising approach for building cost-effective static analyzers. The ideal static analyzer should apply accurate but costly techniques only when they ben...
详细信息
Context: Recently data-driven program analysis has emerged as a promising approach for building cost-effective static analyzers. The ideal static analyzer should apply accurate but costly techniques only when they benefit. However, designing such a strategy for real-world programs is highly nontrivial and requires labor-intensive work. The goal of data-driven program analysis is to automate this process by learning the strategy from data through a learning algorithm. Objective: Current learning algorithms for data-driven program analysis are not scalable enough to be used with large codebases. The objective of this paper is to overcome this shortcoming and present a new algorithm that is able to efficiently learn a strategy from large codebases. Method: The key idea is to use an oracle and transform the existing blackbox learning problem into a whitebox one that is much easier to solve. The oracle quantifies the relative importance of each part of the program with respect to the analysis precision. The oracle can be obtained by running the most and least precise analyses only once over the codebase. Results: Our learning algorithm is much faster than the existing algorithms while producing high quality strategies. The evaluation is done with 140 open-source C programs, comprising of 2.1 MLoC in total. learning at this large scale was previously impractical. Conclusion: Our work advances the state-of-the-art of data-driven program analysis by addressing the scalability issue of the existing learning algorithm. Our technique will make the data-driven approach more practical in the real-world.
In the communication aspect of a computer network, data are sent by packets. If the communication channel is not completely safe, then the arrival of the packets must be acknowledged. In the data-acknowledgment proble...
详细信息
In the communication aspect of a computer network, data are sent by packets. If the communication channel is not completely safe, then the arrival of the packets must be acknowledged. In the data-acknowledgment problem, the goal is to determine the time of sending acknowledgments. Here we present a new online algorithm for it, where the algorithm itself is a parameter-learning extension of the alarming algorithms. The efficiency of the algorithm is then investigated by testing it experimentally, and it is demonstrated that the new parameter-learning algorithm performs significantly better than the original one.
FeedForward (FF-) Blind Source Separation (BSS) systems have some degree of freedom in the solution space. Therefore, signal distortion is likely to occur. First, a criterion for the signal distortion is discussed. Pr...
详细信息
FeedForward (FF-) Blind Source Separation (BSS) systems have some degree of freedom in the solution space. Therefore, signal distortion is likely to occur. First, a criterion for the signal distortion is discussed. Properties of conventional methods proposed to suppress the signal distortion are analyzed. Next, a general condition for complete separation and distortion-free is derived for multi-channel FF-BSS systems. This condition is incorporated in learning algorithms as a distortion-free constraint. Computer simulations using speech signals and stationary colored signals are performed for the conventional methods and for the new learning algorithms employing the proposed distortion-free constraint. The proposed method can well suppress signal distortion, while maintaining a high source separation performance.
A new on-line direct control scheme for the Autonomous Underwater Vehicles (AUV), using recurrent neural networks, is investigated. In the proposed scheme, the controller consists of a three-layer network architecture...
详细信息
A new on-line direct control scheme for the Autonomous Underwater Vehicles (AUV), using recurrent neural networks, is investigated. In the proposed scheme, the controller consists of a three-layer network architecture having feedforward input and output layers, and a totally recurrent hidden layer All the interconnection strengths are synchronously updated using a computationally inexpensive learning algorithm called Alopex. The updating is based on the output error of the system directly, rather than using a transformed version of the error employed in the other neural network based direct control schemes. In the present implementation, the network starts from random initial conditions without needing any prior training, and learns the dynamics of the AUV to provide the correct control signal. Based on the simulation experiments using the nonlinear dynamics of an AUV, we demonstrate that the proposed learning algorithm and the network architecture provide stable and accurate tracking performance. We have also addressed the issue of robustness of the controller to system parameter variations as well as to measurement disturbances.
This paper proposes a concurrent learning algorithm for forward and inverse modeling. The algorithm is consisted of two phases. In the first phase, a feedback controller is used. The forward model is trained using the...
详细信息
This paper proposes a concurrent learning algorithm for forward and inverse modeling. The algorithm is consisted of two phases. In the first phase, a feedback controller is used. The forward model is trained using the output values of the controller as the input values to the system and the inverse model is trained by the feedback error learning. In the second phase, the forward model and the inverse model are trained at the same time. By the simulation experiments in a two-link manipulator, it is confirmed that our algorithm can converge faster than the ones already proposed.
暂无评论