Reduction Models in Competitive Learning Founded on Distortion Standards

2008 ◽  
Vol 12 (3) ◽  
pp. 314-323 ◽  
Author(s):  
Michiharu Maeda ◽  
◽  
Noritaka Shigei ◽  
Hiromi Miyajima ◽  
Kenichi Suzaki ◽  
...  
Keyword(s):  
Image Coding ◽  
Learning Algorithm ◽  
Computation Time ◽  
Competitive Learning ◽  
Experimental Results ◽  
Error Criterion ◽  
General Reduction ◽  
Predetermined Number

Two reductions in competitive learning founded on distortion standards are discussed from the viewpoint of generating necessary and appropriate reference vectors under the condition of their predetermined number. The first approach is termed the segmental reduction and competitive learning algorithm. The algorithm is presented as follows: First, numerous reference vectors are prepared and the algorithm is processed under competitive learning. Next, reference vectors are sequentially eliminated to reach their prespecified number based on the partition error criterion. The second approach is termed the general reduction and competitive learning algorithm. The algorithm is presented as follows: First, numerous reference vectors are prepared and the algorithm is processed under competitive learning. Next, reference vectors are sequentially erased based on the average distortion criterion. Experimental results demonstrate the effectiveness of our approaches compared to conventional techniques in average distortion. The two approaches are applied to image coding to determine their feasibility in quality and computation time.

COMPENSATION COMPETITIVE LEARNING

2001 ◽  
Vol 01 (03) ◽  
pp. 303-322
Author(s):  
ZHI-QIANG LIU ◽  
YAJUN ZHANG
Keyword(s):  
Input Data ◽  
Learning Algorithm ◽  
A Priori ◽  
Competitive Learning ◽  
Experimental Results ◽  
Initial Number ◽  
Number Of Clusters ◽  
And Performance ◽  
Competitive Algorithms

In general, in competitive learning the requirement for the initial number of prototypes is a difficult task, as we do not usually know the number of clusters in the input data a priori. The behavior and performance of the competitive algorithms are very sensitive to the initial locations and number of the prototypes. In this paper after investigating several important competitive learning paradigms, we present compensation techniques for overcoming the problems in competitive learning. Our experimental results show that competition with compensation can improve the performance of the learning algorithm.

A novel fuzzy entropy-constrained competitive learning algorithm for image coding

2001 ◽  
Vol 37 (1-4) ◽  
pp. 197-208
Author(s):  
Wen-Jyi Hwang ◽  
Faa-Jeng Lin ◽  
Shi-Chiang Liao ◽  
Jeng-Hsin Huang
Keyword(s):  
Image Coding ◽  
Learning Algorithm ◽  
Competitive Learning ◽  
Fuzzy Entropy

Adaptive Vector Quantization with Creation and Reduction Grounded in the Equinumber Principle

2005 ◽  
Vol 9 (6) ◽  
pp. 599-606 ◽  
Author(s):  
Michiharu Maeda ◽  
◽  
Noritaka Shigei ◽  
Hiromi Miyajima ◽  
Keyword(s):  
Neural Networks ◽  
Vector Quantization ◽  
Learning Process ◽  
Image Coding ◽  
Image Data ◽  
Experimental Results ◽  
Structural Learning ◽  
Adaptive Vector Quantization ◽  
Partition Space ◽  
Predetermined Number

This paper concerns the constitution of unit structures in neural networks for adaptive vector quantization. Partition errors are mutually equivalent when the number of inputs in a partition space is mutually equal, and average distortion is asymptotically minimized. This is termed the equinumber principle, in which two types of adaptive vector quantization are presented to avoid the initial dependence of reference vectors. Conventional techniques, such as structural learning with forgetting, have the same number of output units from start to finish. Our approach explicitly changes the number of output units to reach a predetermined number without neighboring relations equalling the numbers of inputs in a partition space. First, output units are sequentially created based on the equinumber principle in the learning process. Second, output units are sequentially deleted to reach a prespecified number. Experimental results demonstrate the effectiveness of these techniques in average distortion. These approaches are applied to image data and their feasibility was confirmed in image coding.

Comparison between an Advanced Numerical Simulation of Sheet Incremental Forming Using Adaptive Remeshing and Experimental Results

2013 ◽  
Vol 554-557 ◽  
pp. 1375-1381 ◽  
Author(s):  
Laurence Giraud-Moreau ◽  
Abel Cherouat ◽  
Jie Zhang ◽  
Houman Borouchaki
Keyword(s):  
Numerical Simulation ◽  
Numerical Model ◽  
Sheet Metal ◽  
Metal Forming ◽  
Tool Path ◽  
Incremental Forming ◽  
Computation Time ◽  
Experimental Results ◽  
Forming Process ◽  
Adaptive Remeshing

Recently, new sheet metal forming technique, incremental forming has been introduced. It is based on using a single spherical tool, which is moved along CNC controlled tool path. During the incremental forming process, the sheet blank is fixed in sheet holder. The tool follows a certain tool path and progressively deforms the sheet. Nowadays, numerical simulations of metal forming are widely used by industry to predict the geometry of the part, stresses and strain during the forming process. Because incremental forming is a dieless process, it is perfectly suited for prototyping and small volume production [1, 2]. On the other hand, this process is very slow and therefore it can only be used when a slow series production is required. As the sheet incremental forming process is an emerging process which has a high industrial interest, scientific efforts are required in order to optimize the process and to increase the knowledge of this process through experimental studies and the development of accurate simulation models. In this paper, a comparison between numerical simulation and experimental results is realized in order to assess the suitability of the numerical model. The experimental investigation is realized using a three-axis CNC milling machine. The forming tool consists in a cylindrical rotating punch with a hemispherical head. A subroutine has been developed to describe the tool path from CAM procedure. A numerical model has been developed to simulate the sheet incremental forming process. The finite element code Abaqus explicit has been used. The simulation of the incremental forming process stays a complex task and the computation time is often prohibitive for many reasons. During this simulation, the blank is deformed by a sequence of small increments that requires many numerical increments to be performed. Moreover, the size of the tool diameter is generally very small compared to the size of the metal sheet and thus the contact zone between the tool and the sheet is limited. As the tool deforms almost every part of the sheet, small elements are required everywhere in the sheet resulting in a very high computation time. In this paper, an adaptive remeshing method has been used to simulate the incremental forming process. This strategy, based on adaptive refinement and coarsening procedures avoids having an initially fine mesh, resulting in an enormous computing time. Experiments have been carried out using aluminum alloy sheets. The final geometrical shape and the thickness profile have been measured and compared with the numerical results. These measurements have allowed validating the proposed numerical model. References [1] M. Yamashita, M. Grotoh, S.-Y. Atsumi, Numerical simulation of incremental forming of sheet metal, J. Processing Technology, No. 199 (2008), p. 163 172. [2] C. Henrard, A.M. Hbraken, A. Szekeres, J.R. Duflou, S. He, P. Van Houtte, Comparison of FEM Simulations for the Incremental Forming Process, Advanced Materials Research, 6-8 (2005), p. 533-542.

A Study of the Continuous Optimization Problem Using a Wood Robot Controlled by a Biologically Motivated System

2015 ◽  
Vol 137 (7) ◽  
Author(s):  
Jong-Chen Chen
Keyword(s):  
Physical Environment ◽  
Intelligent System ◽  
Optimization Problems ◽  
Learning Algorithm ◽  
Continuous Optimization ◽  
Physical Structure ◽  
Experimental Results ◽  
Self Organized ◽  
Selection Operators ◽  
Continuous Optimization Problems

Continuous optimization plays an increasingly significant role in everyday decision-making situations. Our group had previously developed a multilevel system called the artificial neuromolecular system (ANM) that possessed structure richness allowing variation and/or selection operators to act on it in order to generate a broad range of dynamic behaviors. In this paper, we used the ANM system to control the motions of a wooden walking robot named Miky. The robot was used to investigate the ANM system's capability to deal with continuous optimization problems through self-organized learning. Evolutionary learning algorithm was used to train the system and generate appropriate control. The experimental results showed that Miky was capable of learning in a continued manner in a physical environment. A further experiment was conducted by making some changes to Miky's physical structure in order to observe the system's capability to deal with the change. Detailed analysis of the experimental results showed that Miky responded to the change by appropriately adjusting its leg movements in space and time. The results showed that the ANM system possessed continuous optimization capability in coping with the change. Our findings from the empirical experiments might provide us another dimension of information of how to design an intelligent system comparatively friendlier than the traditional systems in assisting humans to walk.

scholarly journals Remote Judgment Method of Painting Image Style Plagiarism Based on Wireless Network Multitask Learning

2021 ◽  
Vol 2021 ◽  
pp. 1-8
Author(s):  
Zhijun Wang
Keyword(s):  
Wireless Network ◽  
Kernel Function ◽  
Digital Image ◽  
Learning Algorithm ◽  
Deep Level ◽  
Multitask Learning ◽  
Experimental Results ◽  
Spatial Consistency ◽  
Discrimination Rate ◽  
Edge Sampling

Since the artistry of the work cannot be accurately described, the identification of reproducible plagiarism is more difficult. The identification of reproducible plagiarism of digital image works requires in-depth research on the artistry of artistic works. In this paper, a remote judgment method for plagiarism of painting image style based on wireless network multitask learning is proposed. According to this new method, the uncertainty of painting image samples is removed based on multitask learning algorithm edge sampling. The deep-level details of the painting image are extracted through the multitask classification kernel function, and most of the pixels in the image are eliminated. When the clustering density is greater than the judgment threshold, it can be considered that the two images have spatial consistency. It can also be judged based on this that the two images are similar, that is, there is plagiarism in the painting. The experimental results show that the discrimination rate is always close to 100%, the misjudgment rate of plagiarism of painting images has been reduced, and the various indicators in the discrimination process are the lowest, which fully shows that a very satisfactory discrimination result can be obtained.

scholarly journals A study of Grammar Analysis in English Teaching With Deep Learning Algorithm

2020 ◽  
Vol 15 (18) ◽  
pp. 20
Author(s):  
Gang Zhang
Keyword(s):  
Deep Learning ◽  
Learning Algorithm ◽  
Experimental Results ◽  
P Value ◽  
Analysis Method ◽  
English Teaching ◽  
Deep Learning Algorithm ◽  
R Value ◽  
Grammar Analysis ◽  
Teaching Grammar

In English teaching, grammar is a very important part. Based on the seq2seq model, a grammar analysis method combining the attention mechanism, word embedding and CNN seq2seq was designed using the deep learning algorithm, then the algorithm training was completed on NUCLE, and it was tested on CoNIL-2014. The experimental results showed that of seq2seq+attention improved 33.43% compared to the basic seq2seq; in the comparison between the method proposed in this study and CAMB, the P value of the former was 59.33% larger than that of CAMB, the R value was 8.9% larger, and the value of was 42.91% larger. Finally, in the analysis of the actual students' grammar homework, the proposed method also showed a good performance. The experimental results show that the method designed in this study is effective in grammar analysis and can be applied and popularized in actual English teaching.

AN IMPROVED EMBEDDED ZEROTREE WAVELET IMAGE CODING METHOD BASED ON COEFFICIENT PARTITIONING USING MORPHOLOGICAL OPERATION

2000 ◽  
Vol 14 (06) ◽  
pp. 795-807 ◽  
Author(s):  
JUNMEI ZHONG ◽  
C. H. LEUNG ◽  
Y. Y. TANG
Keyword(s):  
Image Compression ◽  
Video Coding ◽  
Image Coding ◽  
Experimental Results ◽  
Morphological Operation ◽  
Compression Algorithms ◽  
Still Image ◽  
Coding Method ◽  
Wavelet Image ◽  
Image Compression Algorithm

In recent years, wavelets have attracted great attention in both still image compression and video coding, and several novel wavelet-based image compression algorithms have been developed so far, one of which is Shapiro's embedded zerotree wavelet (EZW) image compression algorithm. However, there are still some deficiencies in this algorithm. In this paper, after the analysis of the deficiency in EZW, a new algorithm based on quantized coefficient partitioning using morphological operation is proposed. Instead of encoding the coefficients in each subband line-by-line, regions in which most of the quantized coefficients are significant are extracted by morphological dilation and encoded first. This is followed by using zerotrees to encode the remaining space which has mostly zeros. Experimental results show that the proposed algorithm is not only superior to the EZW, but also compares favorably with the most efficient wavelet-based image compression algorithms reported so far.

Sign in / Sign up

Export Citation Format

Share Document