Alternative basis matrix multiplication algorithms are the fastest matrix multiplication algorithms in practice to date. However, are they numerically stable?We obtain the first numerical error bound for alternative b...
详细信息
ISBN:
(数字)9798350387117
ISBN:
(纸本)9798350387124
Alternative basis matrix multiplication algorithms are the fastest matrix multiplication algorithms in practice to date. However, are they numerically stable?We obtain the first numerical error bound for alternative basis matrix multiplication algorithms, demonstrating that their error bounds are asymptotically identical to the standard fast matrix multiplication algorithms, such as Strassen’s. We further show that arithmetic costs and error bounds of alternative basis algorithms can be simultaneously and independently optimized. Particularly, we obtain the first fast matrix multiplication algorithm with a 2-by-2 base case that simultaneously attains the optimal leading coefficient for arithmetic costs and optimal asymptotic error bound, effectively beating the Bini and Lotti (1980) speed-stability trade-off for fast matrix multiplication. We provide high-performance parallel implementations of our algorithms with benchmarks that show our algorithm is on par with the best in class for speed and with the best in class for stability. Finally, we show that diagonal scaling stability improvement techniques for fast matrix multiplication are as effective for alternative basis algorithms, both theoretically and empirically. These findings promote the use of alternative basis matrix multiplication algorithms in practical applications.
TSignatures are one of the most frequently used biometric authentication systems, they cannot be stolen or loaned to others. Signature recognition can be done by eye, but signature identification using this method can...
详细信息
ISBN:
(纸本)9781665486231
TSignatures are one of the most frequently used biometric authentication systems, they cannot be stolen or loaned to others. Signature recognition can be done by eye, but signature identification using this method can be fooled by someone's exper-tise in forging signatures. The absence of special characteristics of a signature that makes it difficult for people to forge it. Human physical limitations (fatigue, inaccuracy and impaired vision) can affect the interpretation of a signature. Therefore, automatic signature identification can assist forensic experts in processing and analyzing signatures. This study compares the backpropagation and radial basis artificial neural network methods. The data used are 96 images, consisting of 84 training images and 12 test images. Feature extraction used is Fast Fourier Transform, which is scanned horizontally and vertically. The backpropagation architecture uses traingdx learning, the variable rate of learning is 0.1, the ratio of increasing the learning rate is 1.6, the ratio is decreasing the learning rate is 0.5, and the momentum is 0.3. The first hidden layer is 48, the second hidden layer is 24 and the output is 12. The identification accuracy using the backpropagation method is 91.67%. The architecture of the radial basis method uses a data spread width of 1, epoh 10.000, goal 1e-2. The identification accuracy using the radial basis method is 66.67%. This research proves the backpropagation method is more accurate than the radial basis, if it is used to identify biometric systems with signature specimens.
The algorithm of whitening which can be used during the initial stage of functioning of the cryptographic system of information security is obtained in the work. The algorithm allows to remove stochastic relations of ...
详细信息
ISBN:
(数字)9781728155661
ISBN:
(纸本)9781728155678
The algorithm of whitening which can be used during the initial stage of functioning of the cryptographic system of information security is obtained in the work. The algorithm allows to remove stochastic relations of a source message which will make it harder to crack it by intruders. Non-linear canonical expansion of random sequences forms the basis of the algorithm. The results of the experimental investigations show high efficiency of the offered whitening method.
In this paper we discuss an algorithm to determine the level of attention which an aircraft under control requires in a sectorless ATM environment. Data were gathered by means of an experiment in which air traffic con...
详细信息
ISBN:
(数字)9781624105784
ISBN:
(纸本)9781624105784
In this paper we discuss an algorithm to determine the level of attention which an aircraft under control requires in a sectorless ATM environment. Data were gathered by means of an experiment in which air traffic controllers from Hungarocontrol were tasked with evaluating scenarios within a sectorless airspace. These controllers were asked how much attention an aircraft under control requires, as well as the influence of the surrounding traffic on it. With this data a set of rules was developed, using the principles of fuzzy logic, which will form the basis of the algorithm. On the basis of this algorithm several tools for the sectorless ATM concept can be developed like a priority list of the aircraft under control, a filtering algorithm for the radar display and a heat map showing expected critical points on the trajectory of an aircraft.
Many modern data analysis tasks often require one to efficiently handle and analyze large matrix-form datasets such as term-document matrices and spatiotemporal measurements made via sensor networks. Since such matric...
详细信息
ISBN:
(纸本)9781509007479
Many modern data analysis tasks often require one to efficiently handle and analyze large matrix-form datasets such as term-document matrices and spatiotemporal measurements made via sensor networks. Since such matrices are often shuffled and scrambled, they do not have spatial coherency and smoothness that usual images and photographs possess, and consequently, the conventional wavelets and their relatives cannot be used in practice. Instead we propose to use our multiscale basis dictionaries for graphs, i.e., the Generalized Haar-Walsh Transform. In particular, we build such dictionaries for columns and rows separately, extract the column best basis and the row best basis from the basis dictionaries, and construct the tensor product of such best bases, which turns out to reveal hidden dependency and underlying geometric structure in the given matrix data. Finally, we will demonstrate the effectiveness of our approach using the Science News database.
The aim of this study is to compare the performance assessment results of the different classification methods and ensemble algorithms for the detection of chronic kidney disease. Six different basic classifier (naive...
详细信息
The aim of this study is to compare the performance assessment results of the different classification methods and ensemble algorithms for the detection of chronic kidney disease. Six different basic classifier (naive bayes, k nearest neighbor (KNN), support vector machines (SVM), J48, random trees, decision tables) and three different ensemble algorithm (adaboost, bagging, random subspace) are used in the study. Classification results were evaluated using three different performance evaluation criteria (accuracy, kappa, the area under the ROC curve (AUC)). According to the performance evaluation results, J48 basis algorithm for use with bagging and random subspace ensemble algorithms and random tree basis algorithm for use with bagging ensemble algorithm has provided 100% classification success.
Given a 0-dimensional polynomial system in a polynomial ring over F_2 having only F_2-rational solutions, we optimize the Border basis Algorithm (BBA) for solving this system by introducing a Boolean BBA. This algorit...
详细信息
ISBN:
(纸本)9781509057085
Given a 0-dimensional polynomial system in a polynomial ring over F_2 having only F_2-rational solutions, we optimize the Border basis Algorithm (BBA) for solving this system by introducing a Boolean BBA. This algorithm is further improved by optimizing the linear algebra steps. We discuss ways to combine it with SAT solvers, optimized methods for performing the combinatorial steps involved in the algorithm, and various approaches to implement the linear algebra steps. Based on our C++ implementation, we provide some timings to compare sparse and dense representations of the coefficient matrices and to Gröebner basis methods.
Heart rate can be calculated by Photoplethysmographic (PPG) signal which is one of the most important biological signals in the field of wearable monitoring. But motion artifacts and noise artifacts have such a bad in...
详细信息
Heart rate can be calculated by Photoplethysmographic (PPG) signal which is one of the most important biological signals in the field of wearable monitoring. But motion artifacts and noise artifacts have such a bad influence on the quality of PPG signal that it is hard to extract pure heart rate. In this paper, a novel Motion and Noise Artifacts Reduction Mechanism (MNARM) is presented to remove motion artifacts and noise artifacts of PPG signals and calculate heart rate in real time based on a wireless, wrist wearable PPG device. MNARM combines NLMS (Normalized Least Mean Square) adaptive filtering algorithm and Mallat algorithm to remove artifacts which is on the different frequency range with heart rate and artifacts which is not respectively. The novelty of MNARM lied in two facts: a 6-axis acceleration signals are employed as the reference signal; and the combination of the two fundamental algorithms reduces the complexity and the amount of calculation. Therefore it is more suitable and effective for the wearable device to extract heart rate. In experimental tests, results indicate that artifacts are greatly eliminated, the device based on MNARM can measure accurate heart rate in real time in the case of subjects' gender and age vary or exercise intensity varies in limited range.
The existing multi-tone jamming suppression algorithms for direct sequence spread spectrum (DSSS) communications are confined to the high sampling rate. The compressive sensing (CS) was applied to solve the problem. F...
详细信息
ISBN:
(纸本)9781467390996
The existing multi-tone jamming suppression algorithms for direct sequence spread spectrum (DSSS) communications are confined to the high sampling rate. The compressive sensing (CS) was applied to solve the problem. Firstly, the DSSS signal and multi-tone jamming sparse dictionary was built, the multi-tone jamming suppression algorithm used in compressed domain was designed. Secondly, due to the difficulty in getting the prior information of the sparse degree of the jamming, the adaptive multi-tone jamming suppression algorithm was designed by setting the control threshold. The theoretical basis of the algorithm were analyzed, the effectiveness of the algorithm was verified by computer simulation. The results show that the algorithm could suppress the multi-tone jamming effectively, the performance don't change over the jamming intensity, jamming quantity and jamming position. This will provide an effective method for the reconstruction of the compressed DSSS signal in the condition of the multi-tone jamming.
In this paper, an improved music genre classification method is presented. The proposed method makes use of the wavelet package transform (WPT) and the best basis algorithm (BBA) to accurately classify and increase cl...
详细信息
ISBN:
(纸本)9781467302197
In this paper, an improved music genre classification method is presented. The proposed method makes use of the wavelet package transform (WPT) and the best basis algorithm (BBA) to accurately classify and increase classification performance. It is well known that WPT can generate a wavelet decomposition that offers a richer signal analysis. In this paper, the music signal is first decomposed into approximation and detail coefficients using WPT with the best basis algorithm to minimize the Shannon entropy and maximize the representation of music signal. This paper uses the Top-Down search strategy with cost function to select the best basis. Then the proposed method could apply support vector machine (SVM) to build a music genre classifier using the mel-frequency cepstral coefficients (MFCC) and log energies extracted from the decomposition coefficients of WPT with the best basis algorithm. Finally one can perform music genre classification with the built music genre classifier. Experiments conducted on three different music datasets have shown that the proposed method can achieve higher classification accuracy than other music genre classification methods with the same experimental setup.
暂无评论