In the context of accurately measuring harmonic contributions, it is essential to assess both the overall stable harmonic responsibilities and the transient fluctuations to determine if short-term harmonic emissions e...
详细信息
We propose a highly efficient framework for kernel multi-class models with a large and structured set of classes. Kernel parameters are learned automatically by maximizing the cross-validation log likelihood, and pred...
详细信息
ISBN:
(纸本)9780262195683
We propose a highly efficient framework for kernel multi-class models with a large and structured set of classes. Kernel parameters are learned automatically by maximizing the cross-validation log likelihood, and predictive probabilities are estimated. We demonstrate our approach on large scale text classification tasks with hierarchical class structure, achieving state-of-the-art results in an order of magnitude less time than previous work.
Many machine learning tasks can be formulated as a stochastic compositional optimization (SCO) problem such as reinforcement learning, AUC maximization and meta-learning, where the objective function involves a nested...
详细信息
Many machine learning tasks can be formulated as a stochastic compositional optimization (SCO) problem such as reinforcement learning, AUC maximization and meta-learning, where the objective function involves a nested composition associated with an expectation. Although many studies have been devoted to studying the convergence behavior of SCO algorithms, there is little work on understanding their generalization, that is, how these learning algorithms built from training data would behave on future test examples. In this paper, we provide the stability and generalization analysis of stochastic compositional gradient descent algorithms in the framework of statistical learning theory. Firstly, we introduce a stability concept called compositional uniform stability and establish its quantitative relation with generalization for SCO problems. Then, we establish the compositional uniform stability results for two notable stochastic compositional gradient descent algorithms, namely SCGD and SCSC. Finally, we derive dimension-independent excess risk bounds for SCGD and SCSC by balancing stability results and optimization errors. To the best of our knowledge, these are the first-ever known results on stability and generalization analysis of stochastic compositional gradient descent algorithms. Copyright 2024 by the author(s)
In this paper, we propose an analytical framework to statistically analyze the battery recharging time (BRT) in reconfigurable intelligent surfaces (RISs)-assisted wireless power transfer (WPT) systems. Specifically, ...
详细信息
A computationally efficient method of three-dimensional (3-D) localization of underwater acoustic sources using an acoustic vector sensor (AVS) array is presented in this paper. Noise is modeled as a Gaussian mixture ...
详细信息
ISBN:
(纸本)9789380095455
A computationally efficient method of three-dimensional (3-D) localization of underwater acoustic sources using an acoustic vector sensor (AVS) array is presented in this paper. Noise is modeled as a Gaussian mixture (GM), and an expectationmaximization (EM) algorithm is used to estimate all the unknown parameters. The proposed method consists of an initialization stage followed by iterative updates which converge in a small number of iterations. Performance of the proposed method is better than that of existing methods, and performance of the AVS array is much better than that of the acoustic pressure sensor (APS) array.
Automatic Modulation Classification (AMC) is extremely desirable to realize the merits of cognitive radios in plethora of commercial and military applications. However, AMC is extremely challenging in real-world scena...
详细信息
ISBN:
(纸本)9781479946532
Automatic Modulation Classification (AMC) is extremely desirable to realize the merits of cognitive radios in plethora of commercial and military applications. However, AMC is extremely challenging in real-world scenarios due to multipath fading and additive Gaussian noise on modulation schemes. Moreover, it becomes more difficult in blind environments where a little or no priori information about the received signal is available. Most of the available modulation classifiers do not consider the fading effects which results in performance degradation of classification in a blind channel environment. In this paper, we investigate the multipath fading effects on the AMC of some common modulation schemes i.e., BPSK, QPSK and 16-QAM for blind channels. In our work channel is assumed to be suffering from multipath and excessive additive noise resulting in low SNR of signal. The unknown channel and noise parameters are estimated using Hidden Markovian based expectation maximization algorithm. The estimated channel coefficients are then used in Maximum-Likelihood classifier for the classification of modulation scheme. Simulation results show that phase modulation schemes (i.e., BPSK and QPSK) perform better at low SNR compared to 16-QAM which includes the phase amplitude information.
In order to investigate the performance of visual feature extraction method for automatic image annotation, three visual feature extraction methods, namely discrete cosine transform, Gabor transform and discrete wavel...
详细信息
ISBN:
(纸本)9781424427932
In order to investigate the performance of visual feature extraction method for automatic image annotation, three visual feature extraction methods, namely discrete cosine transform, Gabor transform and discrete wavelet transform, are studied in this paper. These three methods are used to extract low-level visual feature vectors from images in a given database separately, then these feature vectors are mapped to high-level semantic words to annotate images with labels in a given semantic label set. As it is more efficient to depict the visual features of an image by the feature distribution than to resort to image segmentation technology for semantic image blocks, this paper is going to find out which of the three feature extraction methods performs better in image annotation based on the distribution of feature vectors from the image. The performance of three different kinds of feature extraction method is fully analyzed, and it is found that discrete cosine transform method is more suitable for Gaussian mixture model in automatic image annotation.
A target can be positioned by wireless communication sensors. When the range based sensors have biased measurements, an expectationmaximization (EM) algorithm is proposed to jointly estimate the target state and sens...
详细信息
ISBN:
(纸本)9781467391047
A target can be positioned by wireless communication sensors. When the range based sensors have biased measurements, an expectationmaximization (EM) algorithm is proposed to jointly estimate the target state and sensors' biases, including the batch EM and sliding window EM algorithms. To implement the algorithms, the Iterated Extended Kalman Smoother (IEKS) is also embedded in the EM algorithm. The simulation results show that the batch algorithm has the best estimation performance. The sliding window EM algorithm has better estimation performance than the augmented UKF (AUKF) algorithm. Since batch EM algorithm is not suitable for real time estimation scenario, the sliding window EM algorithm is recommended for real time target positioning.
This paper synthesises the in-depth study and extensive practical application of the principles of probabilistic statistics and the Gaussian mixture clustering (GMM) algorithm with the aim of revealing the powerful ef...
详细信息
暂无评论