发布求助
文献互助 智能选刊 最新文献

2008 IEEE Workshop on Machine Learning for Signal Processing最新文献

筛选
英文 中文
F-SVC: A simple and fast training algorithm soft margin Support Vector Classification F-SVC:一种简单快速的软边界支持向量分类训练算法
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685503
M. Tohmé, R. Lengellé
{"title":"F-SVC: A simple and fast training algorithm soft margin Support Vector Classification","authors":"M. Tohmé, R. Lengellé","doi":"10.1109/MLSP.2008.4685503","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685503","url":null,"abstract":"Support vector machines have obtained much success in machine learning. But their training require to solve a quadratic optimization problem so that training time increases dramatically with the increase of the training set size. Hence, standard SVM have difficulty in handling large scale problems. In this paper, we present a new fast training algorithm for soft margin support vector classification. This algorithm searches for successive efficient feasible directions. A heuristic for searching the direction maximally correlated with the gradient is applied and the optimum step size of the optimization algorithm is analytically determined. Furthermore the solution, gradient and objective function are recursively obtained. In order to deal with large scale problems, the Gram matrix has not to be stored. Our iterative algorithm fully exploits quadratic functions properties. F-SVC is very simple, easy to implement and able to perform on large data sets.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115364884","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 5
Multiple TDOA estimation by using a state coherence transform for solving the permutation problem in frequency-domain BSS 基于状态相干变换的多重TDOA估计解决频域BSS中的排列问题
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685453
F. Nesta, M. Omologo, P. Svaizer
{"title":"Multiple TDOA estimation by using a state coherence transform for solving the permutation problem in frequency-domain BSS","authors":"F. Nesta, M. Omologo, P. Svaizer","doi":"10.1109/MLSP.2008.4685453","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685453","url":null,"abstract":"A novel method to solve the permutation problem for Blind Source Separation (BSS) is presented. According to the acoustic propagation model, in frequency-domain, each separation matrix can be represented with a set of states associated with each source. We formulate a novel transform of the states which is independent of the aliasing and of the permutations since states belonging to all the sources are exploited at the same time. The estimated TDOAs are used to model the propagation of the acoustic wave and to cluster all the frequency components associated to the same source. Experimental results show that the novel approach can be applied to localize and separate sources in challenging situations: two sources have been separated estimating long demixing filters (0.25-0.5s) using widely spaced microphones (0.25 m) in reverberant environment (T60 = 700 ms) and using very short signals (0.5-1 s).","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121224350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 18
Gene expression dissection by non-negative well-grounded source separation 非负性良好接地源分离的基因表达分离
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685489
Yitan Zhu, Tsung-Han Chan, E. Hoffman, Y. Wang
{"title":"Gene expression dissection by non-negative well-grounded source separation","authors":"Yitan Zhu, Tsung-Han Chan, E. Hoffman, Y. Wang","doi":"10.1109/MLSP.2008.4685489","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685489","url":null,"abstract":"A linear mixture model of non-negative sources is used to dissect the gene expression data into components that are putative underlying active biological processes. Each biological process/component is characterized by its specific genes that are exclusively highly expressed in it and expected to be functional enriched; while a majority of all the genes maintain basic cellular structure and functions to support these specific genes and thus are roughly commonly expressed across all components. Such components form non-negative well-grounded, but dependent and non-sparse sources in the model. The unique identifiability of the model is proved. A blind source separation method utilizing convex analysis and sector-based clustering is developed with stability analysis based model order selection scheme to identify the components and their activity curves. When applied on muscle regeneration data, our method revealed four underlying active biological processes associated with four successive phases in muscle regeneration.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125347132","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 4
Synchronization and comparison of Lifelog audio recordings 同步和比较的生活日志音频记录
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685526
Andreas Brinch Nielsen, Lars Kai Hansen
{"title":"Synchronization and comparison of Lifelog audio recordings","authors":"Andreas Brinch Nielsen, Lars Kai Hansen","doi":"10.1109/MLSP.2008.4685526","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685526","url":null,"abstract":"We investigate concurrent dasiaLifelogpsila audio recordings to locate segments from the same environment. We compare two techniques earlier proposed for pattern recognition in extended audio recordings, namely cross-correlation and a fingerprinting technique. If successful, such alignment can be used as a preprocessing step to select and synchronize recordings before further processing. The two methods perform similarly in classification, but fingerprinting scales better with the number of recordings, while cross-correlation can offer sample resolution synchronization. We propose and investigate the benefits of combining the two. In particular we show that the combination allows sample resolution synchronization and scalability.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115574806","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 2
Vocal tract area based artificial bandwidth extension 基于声道区域的人工带宽扩展
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685527
K. Kalgaonkar, M. Clements
{"title":"Vocal tract area based artificial bandwidth extension","authors":"K. Kalgaonkar, M. Clements","doi":"10.1109/MLSP.2008.4685527","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685527","url":null,"abstract":"In this paper we present a method of artificial bandwidth extension based on vocal tract areas. This method exploits the many-to-one mapping between the vocal tract areas and linear prediction polynomial, and the quasi-stationary nature of speech to estimate the high-frequency (4 kHz to 8 kHz) components of speech given the narrowband signal (0 to 4 kHZ). Contrary to some of the pre-existing methods, this algorithm does not estimate the missing high-frequency components but directly estimates the broadband vocal tract areas using codebook mapping and ergodic HMMs, to synthesize broadband speech.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128371642","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 4
Learning on varifolds 学习变量
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685510
L. Ding
{"title":"Learning on varifolds","authors":"L. Ding","doi":"10.1109/MLSP.2008.4685510","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685510","url":null,"abstract":"In this paper, we propose a new learning framework based on the mathematical concept of varifolds (Morgan, 2000), which are the measure-theoretic generalization of differentiable manifolds. We compare varifold learning with the popular manifold learning and demonstrate some of its specialties. Algorithmically, we derive a neighborhood refinement technique for hypergraph models, which is conceptually analogous to varifolds, give the procedure for constructing such hypergraphs from data and finally by using the hypergraph Laplacian matrix we are able to solve high-dimensional classification problems accurately.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124334352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 1
Digital circuit design of ICA based implementation of FPGA for real time Blind Signal Separation 基于ICA的数字电路设计，FPGA实现实时盲信号分离
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685476
M. Ounas, S. Chitroub, R. Touhami, M. Yagoub
{"title":"Digital circuit design of ICA based implementation of FPGA for real time Blind Signal Separation","authors":"M. Ounas, S. Chitroub, R. Touhami, M. Yagoub","doi":"10.1109/MLSP.2008.4685476","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685476","url":null,"abstract":"The application of independent component analysis (ICA) algorithm can achieve a real time blind signal separation (BSS) if it is physically implemented using hardware devices. However, due principally to both of the limited size and of the microelectronics technology of the used hardware devices, many practical problem can be encountered to reach the real time processing since the application of the ICA algorithm requires the consumption of a huge number of input signal samples. Hence, the system performance was degraded since we required the consumption of an important number of memory circuits with faster hardware execution time. Therefore, in order to improve the hardware performances of the device, in this paper, the authors proposed the sequential processing of one neuron hardware model based on field programmable gate array (FPGA) implementation. Such approach overcomes the interconnections complexities of the FPGA architecture. Thus, an optimal digital circuit design can be proposed to avoid the consumption of much hardware resources in which a maximum number of samples can be handled while maintaining high speed of hardware processing time. The proposed approach was demonstrated through the experimental study of TIMIT data base exhibiting a hardware execution time of 3.3 mus to process 10000 samples with 57 KHz of sample rates to separate two output independent signals from two input mixed signals.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124438177","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 4
Blind separation of two multi-level sources from a single linear mixture 从单一线性混合中盲分离两个多级源
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685457
K. Diamantaras
{"title":"Blind separation of two multi-level sources from a single linear mixture","authors":"K. Diamantaras","doi":"10.1109/MLSP.2008.4685457","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685457","url":null,"abstract":"We present a novel method for the blind separation of two multi-level sources from one observed instantaneous linear mixture. Our approach is based on the geometric characteristics of the cluster of points and in particular in the distribution of the distances between the cluster centers. In the noiseless case the core of the method is conceptually very simple, deterministic, non-iterative and fast. In the noisy case, the core-algorithm must be preceded by a clustering scheme, for example, the EM algorithm in order to identify the centers. The method works directly on the time domain and so no signal transformation is required.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"160 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131445936","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 9
Fusion of heterogeneous data sources: A quaternionic approach 异构数据源的融合:四元数方法
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685523
C. C. Took, Danilo P. Mandic
{"title":"Fusion of heterogeneous data sources: A quaternionic approach","authors":"C. C. Took, Danilo P. Mandic","doi":"10.1109/MLSP.2008.4685523","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685523","url":null,"abstract":"Sequential fusion of three- and four- dimensional heterogeneous data is achieved in the quaternion space H. This way, data from multiple sensors are combined in order to achieve ldquoimproved accuraciesrdquo and more specific inferences that could not be performed by the use of only a single sensor. To this end, the quaternion LMS (QLMS) is proposed for the online fusion of hypercomplex data within the ldquodata fusion via vector spacesrdquo framework. Case studies on real-world signals such as environmental and financial time series are provided to support the proposed approach.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133506428","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 9
Online Bayesian kernel regression from nonlinear mapping of observations 非线性观测值映射的在线贝叶斯核回归
2008 IEEE Workshop on Machine Learning for Signal Processing Pub Date : 2008-11-21 DOI: 10.1109/MLSP.2008.4685498
M. Geist, O. Pietquin, G. Fricout
{"title":"Online Bayesian kernel regression from nonlinear mapping of observations","authors":"M. Geist, O. Pietquin, G. Fricout","doi":"10.1109/MLSP.2008.4685498","DOIUrl":"https://doi.org/10.1109/MLSP.2008.4685498","url":null,"abstract":"In a large number of applications, engineers have to estimate a function linked to the state of a dynamic system. To do so, a sequence of samples drawn from this unknown function is observed while the system is transiting from state to state and the problem is to generalize these observations to unvisited states. Several solutions can be envisioned among which regressing a family of parameterized functions so as to make it fit at best to the observed samples. However classical methods cannot handle the case where actual samples are not directly observable but only a nonlinear mapping of them is available, which happen when a special sensor has to be used or when solving the Bellman equation in order to control the system. This paper introduces a method based on Bayesian filtering and kernel machines designed to solve the tricky problem at sight. First experimental results are promising.","PeriodicalId":447191,"journal":{"name":"2008 IEEE Workshop on Machine Learning for Signal Processing","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2008-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133594278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 3
下一页 尾页
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整，为了账户安全，请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
请完成安全验证×
相关产品
×
本文献相关产品
联系我们：info@booksci.cn Book学术提供免费学术资源搜索服务，方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号：604180095
Book学术官方微信