作者: 小教堂 時(shí)間: 2025-3-21 20:40
H. R. Koelz,P. G. Lankisch,S. Müller-LissnerTri-Class SVMs. The proposed framework is applied to facial expressions recognition task. The results show that . can exploit effectively the independent views and the unlabeled data to improve the recognition accuracy of facial expressions.作者: placebo 時(shí)間: 2025-3-22 02:58 作者: 大酒杯 時(shí)間: 2025-3-22 07:29 作者: expound 時(shí)間: 2025-3-22 12:38
Electrical Circuits of Ordinary Capacitorsng algorithm which has been found to have numerous advantages over Evolution Strategies. Our empirical results confirm the promise of this approach, and we discuss how it can be scaled up to expert-level Go players.作者: evanescent 時(shí)間: 2025-3-22 15:25 作者: 發(fā)誓放棄 時(shí)間: 2025-3-22 18:35 作者: 調(diào)色板 時(shí)間: 2025-3-23 00:53
Hidden Markov Model for Human Decision Process in a Partially Observable Environment), which incorporates inference of a hidden variable in the environment and switching between exploration and exploitation. Our HMM-based model well reproduced the human behaviors, suggesting the human subjects actually performed exploration and exploitation to effectively adapt to this uncertain environment.作者: BRAWL 時(shí)間: 2025-3-23 03:15
Representing, Learning and Extracting Temporal Knowledge from Neural Networks: A Case Studyis again symbolically represented, incorporating both initial model and learned specification, as shown by our case study. The case study illustrates how the integration of methodologies and principles from distinct AI areas can be relevant to build robust intelligent systems.作者: SUE 時(shí)間: 2025-3-23 08:06
Multi-Dimensional Deep Memory Atari-Go Players for Parameter Exploring Policy Gradientsng algorithm which has been found to have numerous advantages over Evolution Strategies. Our empirical results confirm the promise of this approach, and we discuss how it can be scaled up to expert-level Go players.作者: PARA 時(shí)間: 2025-3-23 13:46
Support Vector Machines-Kernel Algorithms for the Estimation of the Water Supply in Cyprushance the quality of the results and to offer an optimization approach. The final models that were produced have proven to perform with an error of very low magnitude in the testing phase when first time seen data were used.作者: implore 時(shí)間: 2025-3-23 13:57 作者: finite 時(shí)間: 2025-3-23 21:46
An Online Incremental Learning Support Vector Machine for Large-scale Data long testing time. In this paper, we propose an online incremental learning SVM for large data sets. The proposed method mainly consists of two components, Learning Prototypes (LPs) and Learning SVs (LSVs). Experimental results demonstrate that the proposed algorithm is effective for incremental learning problems and large-scale problems.作者: 背帶 時(shí)間: 2025-3-24 01:00 作者: 敬禮 時(shí)間: 2025-3-24 03:23 作者: 公理 時(shí)間: 2025-3-24 08:11
,Good Outcomes from the , (1994–1995),d stage the selected pairs for update often appear repeatedly during the algorithm. Taking advantage of this, we shall propose a procedure to combine previously used descent directions that results in much fewer iterations in this second stage and that may also lead to noticeable savings in kernel operations.作者: Communal 時(shí)間: 2025-3-24 14:29
https://doi.org/10.1007/978-3-662-07212-7 long testing time. In this paper, we propose an online incremental learning SVM for large data sets. The proposed method mainly consists of two components, Learning Prototypes (LPs) and Learning SVs (LSVs). Experimental results demonstrate that the proposed algorithm is effective for incremental learning problems and large-scale problems.作者: Emmenagogue 時(shí)間: 2025-3-24 18:09
https://doi.org/10.1007/978-3-658-34481-8ns. In the metaphor evaluation process, the candidate nouns are evaluated based on the similarities between the meanings of metaphors including the candidate nouns and the meaning of the input expression.作者: anesthesia 時(shí)間: 2025-3-24 19:45 作者: 赤字 時(shí)間: 2025-3-25 02:03 作者: 充氣女 時(shí)間: 2025-3-25 04:34 作者: 蝕刻 時(shí)間: 2025-3-25 10:16
A New Tree Kernel Based on SOM-SDhich adds information about the relative position of subtrees (the route) to the activation of the nodes in such a way to discriminate even those subtrees originally encoded by the same prototypes. Experiments have been performed against two well known benchmark datasets with promising results.作者: Accede 時(shí)間: 2025-3-25 12:49 作者: AGOG 時(shí)間: 2025-3-25 16:15
Breakdown of Thin-Film Dielectricsntification (FDI) of industrial systems [1]. Preparation of experimental conditions in order to collect informative measurements can be very expensive and the data acquired form real-world system may be also very noisy, therefore using all the available data may lead to significant systematic modelling errors.作者: Abduct 時(shí)間: 2025-3-25 21:26 作者: 絆住 時(shí)間: 2025-3-26 03:18
Selection of Training Data for Locally Recurrent Neural Networkntification (FDI) of industrial systems [1]. Preparation of experimental conditions in order to collect informative measurements can be very expensive and the data acquired form real-world system may be also very noisy, therefore using all the available data may lead to significant systematic modelling errors.作者: Hectic 時(shí)間: 2025-3-26 08:18
A Statistical Appraoch to Image Reconstruction from Projections Problem Using Recurrent Neural Netwoomography. The reconstruction process is performed using in this way constructed neural network solving the optimization problem. Computer experiments show that the appropriately designed recurrent neural network is able to reconstruct an image with better quality in comparison to the standart analytical reconstruction algorithm.作者: Binge-Drinking 時(shí)間: 2025-3-26 09:55 作者: 被詛咒的人 時(shí)間: 2025-3-26 13:02 作者: 秘傳 時(shí)間: 2025-3-26 18:32
Conference proceedings 2010ng structure inspired by biolo- cal nervous systems and they consist of a large number of highly interconnected processing elements (neurons). Each neuron is a simple processor with a limited computing capacity typically restricted to a rule for combining input signals (utilizing an activation funct作者: Promotion 時(shí)間: 2025-3-26 21:56
C. Niederau,P. G. Lankisch,S. Müller-Lissnert provide different resolution may be used at the same time, and difficult problems that require highly complex decision borders may be solved in a simple way. Relation of this approach to Support Vector Machines and Liquid State Machines is discussed.作者: Geyser 時(shí)間: 2025-3-27 04:14 作者: arbovirus 時(shí)間: 2025-3-27 06:21
Convergence Improvement of Active Set Training for Support Vector Regressorss paper, we discuss convergence improvement by modifying the training method. To stabilize convergence for a large epsilon tube, we calculate the bias term according to the signs of the previous variables, not the updated variables. And to speed up calculating the inverse matrix by the Cholesky fact作者: outskirts 時(shí)間: 2025-3-27 12:40 作者: legacy 時(shí)間: 2025-3-27 14:57
Support Vector Machines-Kernel Algorithms for the Estimation of the Water Supply in Cyprus the development of an .-Regression Support Vector Machine (SVMR) system with five input parameters. The 5-Fold Cross Validation method was applied in order to produce a more representative training data set. The fuzzy-weighted SVR combined with a fuzzy partition approach was employed in order to en作者: FECK 時(shí)間: 2025-3-27 19:26 作者: reject 時(shí)間: 2025-3-28 01:49 作者: 值得尊敬 時(shí)間: 2025-3-28 04:04
A New Tree Kernel Based on SOM-SDs of methods have their own drawbacks. Kernels typically can only cope with discrete labels and tend to be sparse. On the other side, SOM-SD, an extension of the SOM for structured data, is unsupervised and Markovian, i.e. the representation of a subtree does not consider where the subtree appears i作者: Glucose 時(shí)間: 2025-3-28 06:39
Kernel-Based Learning from Infinite Dimensional 2-Way Tensorshere input data have a natural 2??way representation, such as images or multivariate time series. Our approach aims at relaxing linearity of standard tensor-based analysis while still exploiting the structural information embodied in the input data.作者: 糾纏 時(shí)間: 2025-3-28 12:10 作者: arbiter 時(shí)間: 2025-3-28 14:35 作者: 南極 時(shí)間: 2025-3-28 22:29 作者: 過于平凡 時(shí)間: 2025-3-29 02:14
The Support Feature Machine for Classifying with the Least Number of Featuresnorm of a separating hyperplane. Thus, a classifier with inherent feature selection capabilities is obtained within a single training run. Results on toy examples demonstrate that this method is able to identify relevant features very effectively.作者: Allure 時(shí)間: 2025-3-29 05:51
Hidden Markov Model for Human Decision Process in a Partially Observable Environmentartially observable environment, humans can make appropriate decision by resolving the uncertainty. During decision making in an uncertain environment, resolving behaviors of the uncertainty and optimal behaviors to best suit for the environment are often incompatible, which is termed exploration-ex作者: sleep-spindles 時(shí)間: 2025-3-29 11:13
Representing, Learning and Extracting Temporal Knowledge from Neural Networks: A Case Study Intelligence. Temporal models are fundamental to describe the behaviour of computing and information systems. In addition, acquiring the description of the desired behaviour of a system is a complex task in several AI domains. In this paper, we evaluate a neural framework capable of adapting tempor作者: osteopath 時(shí)間: 2025-3-29 12:21
Multi-Dimensional Deep Memory Atari-Go Players for Parameter Exploring Policy Gradientsame of Go, which, despite its deceivingly simple rules, has eluded the development of artificial expert players. In this paper we attempt to tackle this challenge through a combination of two recent developments in Machine Learning. We employ Multi-Dimensional Recurrent Neural Networks with Long Sho作者: 逃避責(zé)任 時(shí)間: 2025-3-29 17:32
Layered Motion Segmentation with a Competitive Recurrent Networkthat are governed by affine motion patterns. Using an energy-based competitive multilayer architecture based on non-negative activations and multiplicative update rules, we show how the network can perform segmentation tasks that require a combination of affine estimation with local integration and 作者: VOC 時(shí)間: 2025-3-29 23:44 作者: refraction 時(shí)間: 2025-3-30 00:07 作者: 者變 時(shí)間: 2025-3-30 04:06
A Computational System of Metaphor Generation with Evaluation Mechanismrical expression of the form “target (A) like vehicle (B)”. A computational system consisting of a metaphor generation process and a metaphor evaluation process is developed. In the metaphor generation process, a metaphor generation model [1] outputs candidate nouns for vehicles from input expressio作者: Gerontology 時(shí)間: 2025-3-30 11:12 作者: 血統(tǒng) 時(shí)間: 2025-3-30 16:14 作者: defuse 時(shí)間: 2025-3-30 18:28
Lecture Notes in Computer Sciencehttp://image.papertrans.cn/b/image/162698.jpg作者: Blazon 時(shí)間: 2025-3-30 22:32
Kernel-Based Learning from Infinite Dimensional 2-Way Tensorshere input data have a natural 2??way representation, such as images or multivariate time series. Our approach aims at relaxing linearity of standard tensor-based analysis while still exploiting the structural information embodied in the input data.作者: 不愛防注射 時(shí)間: 2025-3-31 02:34 作者: Ordeal 時(shí)間: 2025-3-31 07:09 作者: 調(diào)整校對(duì) 時(shí)間: 2025-3-31 11:16
Layered Motion Segmentation with a Competitive Recurrent Networkthat are governed by affine motion patterns. Using an energy-based competitive multilayer architecture based on non-negative activations and multiplicative update rules, we show how the network can perform segmentation tasks that require a combination of affine estimation with local integration and competition constraints.作者: archetype 時(shí)間: 2025-3-31 14:50
Recurrence Enhances the Spatial Encoding of Static Inputs in Reservoir Networkserefore, we introduce attractor-based reservoir networks for processing of static patterns and compare their performance and encoding capabilities with a related feedforward approach. We show that the network dynamics improve the nonlinear encoding of inputs in the reservoir state which can increase the task-specific performance.作者: hemophilia 時(shí)間: 2025-3-31 19:35
Probability and Its Applicationss paper, we discuss convergence improvement by modifying the training method. To stabilize convergence for a large epsilon tube, we calculate the bias term according to the signs of the previous variables, not the updated variables. And to speed up calculating the inverse matrix by the Cholesky fact作者: CANT 時(shí)間: 2025-3-31 22:21
Invariant Measures and Related Topics,e of their potential interest in applications such as communications. In this work, we focus our attention on the complex gaussian kernel and its possible application in the complex Kernel LMS algorithm. In order to derive the gradients needed to develop the complex kernel LMS (CKLMS), we employ the作者: 鉤針織物 時(shí)間: 2025-4-1 03:17 作者: 戰(zhàn)勝 時(shí)間: 2025-4-1 06:46 作者: constitutional 時(shí)間: 2025-4-1 12:47 作者: 可行 時(shí)間: 2025-4-1 15:10
C. Niederau,P. G. Lankisch,S. Müller-Lissners of methods have their own drawbacks. Kernels typically can only cope with discrete labels and tend to be sparse. On the other side, SOM-SD, an extension of the SOM for structured data, is unsupervised and Markovian, i.e. the representation of a subtree does not consider where the subtree appears i
