Learning articulated motion structures with Bayesian Networks

This paper presents a general methodology for learning articulated motions that, despite having non-linear correlations, are cyclical and have a defined pattern of behavior Using conventional algorithms to extract features from images, a Bayesian classifier is applied to cluster and classify features of the moving object. Clusters are then associated in different frames and structure learning algorithms for Bayesian networks are used to recover the structure of the motion. This framework is applied to the human gait analysis and tracking but applications include any coordinated movement such as multi-robots behavior analysis.

Classifying GABAergic interneurons with semi-supervised projected model-based clustering

Objectives: A recently introduced pragmatic scheme promises to be a useful catalog of interneuron names.We sought to automatically classify digitally reconstructed interneuronal morphologies according tothis scheme. Simultaneously, we sought to discover possible subtypes of these types that might emergeduring automatic classification (clustering). We also investigated which morphometric properties weremost relevant for this classification.Materials and methods: A set of 118 digitally reconstructed interneuronal morphologies classified into thecommon basket (CB), horse-tail (HT), large basket (LB), and Martinotti (MA) interneuron types by 42 of theworld?s leading neuroscientists, quantified by five simple morphometric properties of the axon and fourof the dendrites. We labeled each neuron with the type most commonly assigned to it by the experts. Wethen removed this class information for each type separately, and applied semi-supervised clustering tothose cells (keeping the others? cluster membership fixed), to assess separation from other types and lookfor the formation of new groups (subtypes). We performed this same experiment unlabeling the cells oftwo types at a time, and of half the cells of a single type at a time. The clustering model is a finite mixtureof Gaussians which we adapted for the estimation of local (per-cluster) feature relevance. We performedthe described experiments on three different subsets of the data, formed according to how many expertsagreed on type membership: at least 18 experts (the full data set), at least 21 (73 neurons), and at least26 (47 neurons).Results: Interneurons with more reliable type labels were classified more accurately. We classified HTcells with 100% accuracy, MA cells with 73% accuracy, and CB and LB cells with 56% and 58% accuracy,respectively. We identified three subtypes of the MA type, one subtype of CB and LB types each, andno subtypes of HT (it was a single, homogeneous type). We got maximum (adapted) Silhouette widthand ARI values of 1, 0.83, 0.79, and 0.42, when unlabeling the HT, CB, LB, and MA types, respectively,confirming the quality of the formed cluster solutions. The subtypes identified when unlabeling a singletype also emerged when unlabeling two types at a time, confirming their validity. Axonal morphometricproperties were more relevant that dendritic ones, with the axonal polar histogram length in the [pi, 2pi) angle interval being particularly useful.Conclusions: The applied semi-supervised clustering method can accurately discriminate among CB, HT, LB, and MA interneuron types while discovering potential subtypes, and is therefore useful for neuronal classification. The discovery of potential subtypes suggests that some of these types are more heteroge-neous that previously thought. Finally, axonal variables seem to be more relevant than dendritic ones fordistinguishing among the CB, HT, LB, and MA interneuron types.

Towards improving cluster-based feature selection with a simplified silhouette filter

This paper proposes a filter-based algorithm for feature selection. The filter is based on the partitioning of the set of features into clusters. The number of clusters, and consequently the cardinality of the subset of selected features, is automatically estimated from data. The computational complexity of the proposed algorithm is also investigated. A variant of this filter that considers feature-class correlations is also proposed for classification problems. Empirical results involving ten datasets illustrate the performance of the developed algorithm, which in general has obtained competitive results in terms of classification accuracy when compared to state of the art algorithms that find clusters of features. We show that, if computational efficiency is an important issue, then the proposed filter May be preferred over their counterparts, thus becoming eligible to join a pool of feature selection algorithms to be used in practice. As an additional contribution of this work, a theoretical framework is used to formally analyze some properties of feature selection methods that rely on finding clusters of features. (C) 2011 Elsevier Inc. All rights reserved.

The heterogeneous cluster ensemble method using hubness for clustering text documents

We propose a cluster ensemble method to map the corpus documents into the semantic space embedded in Wikipedia and group them using multiple types of feature space. A heterogeneous cluster ensemble is constructed with multiple types of relations i.e. document-term, document-concept and document-category. A final clustering solution is obtained by exploiting associations between document pairs and hubness of the documents. Empirical analysis with various real data sets reveals that the proposed meth-od outperforms state-of-the-art text clustering approaches.

Segmenting Australian high school students : two-step cluster analysis preliminary results

A one size fits all approach dominates alcohol programs in school settings (Botvin et al., 2007), which may limit program effectiveness (Snyder et al., 2004). Programs tailored to the meet the needs and wants of adolescent groups may be more effective. Limited attention has been directed towards employing a full segmentation process. Where segmentation has been examined, the focus has remained on socio-demographic characteristics and more recently psychographic variables (Mathijssen et al., 2012). The current study aimed to identify whether the addition of behaviour could be used to identify segments. Variables included attitudes towards binge drinking (α = 0.86), behavioral intentions’ (α = 0.97), perceived behavioral control (PBC), injunctive norms (α = 0.94); descriptive norms (α = 0.87), knowledge and reported behaviour. Data was collected from five schools, n = 625 (32.96% girls). Two-Step cluster analysis produced a sample (n = 625) with a silhouette measure of cohesion and separation of 0.4. The intention measure and whether students reported previously consuming alcohol were the most distinguishing characteristics - predictor importance scores of (1.0). A four segment solution emerged. The first segment (“Male abstainers” – 37.2%) featured the highest knowledge score (M: 5.9) along with the lowest-risk drinking attitudes and intentions to drink excessively. Segment 2 (“At risk drinkers” - 11.2%) were characterised by their high-risk attitudes and high-risk drinking intentions. Injunctive (M: 4.1) and descriptive norms (M: 4.9) may indicate a social environment where drinking is the norm. Segment 3 (”Female abstainers” – 25.9%) represents young girls, who have the lowest-risk attitudes and low intentions to drink excessively. The fourth and final segment (boys = 67.4%) (“Moderate drinkers” – 25.7%) all report previously drinking alcohol yet their attitudes and intentions towards excessive alcohol consumption are lower than other segments. Segmentation focuses on identifying groups of individuals who feature similar characteristics. The current study illustrates the importance of including reported behaviour in addition to psychographic and demographic characteristics to identify unique groups to inform intervention planning and design. Key messages The principle of segmentation has received limited attention in the context of school-based alcohol education programs. This research identified four segments amongst 14-16 year high school students, each of which can be targeted with a unique, tailored program to meet the needs and wants of the target audience.

Cluster based training for scaling non-linear support vector machines

Support Vector Machines(SVMs) are hyperplane classifiers defined in a kernel induced feature space. The data size dependent training time complexity of SVMs usually prohibits its use in applications involving more than a few thousands of data points. In this paper we propose a novel kernel based incremental data clustering approach and its use for scaling Non-linear Support Vector Machines to handle large data sets. The clustering method introduced can find cluster abstractions of the training data in a kernel induced feature space. These cluster abstractions are then used for selective sampling based training of Support Vector Machines to reduce the training time without compromising the generalization performance. Experiments done with real world datasets show that this approach gives good generalization performance at reasonable computational expense.

Growth and Modification of Cluster-Assembled Thin Films

Thin film applications have become increasingly important in our search for multifunctional and economically viable technological solutions of the future. Thin film coatings can be used for a multitude of purposes, ranging from a basic enhancement of aesthetic attributes to the addition of a complex surface functionality. Anything from electronic or optical properties, to an increased catalytic or biological activity, can be added or enhanced by the deposition of a thin film, with a thickness of only a few atomic layers at the best, on an already existing surface. Thin films offer both a means of saving in materials and the possibility for improving properties without a critical enlargement of devices. Nanocluster deposition is a promising new method for the growth of structured thin films. Nanoclusters are small aggregates of atoms or molecules, ranging in sizes from only a few nanometers up to several hundreds of nanometers in diameter. Due to their large surface to volume ratio, and the confinement of atoms and electrons in all three dimensions, nanoclusters exhibit a wide variety of exotic properties that differ notably from those of both single atoms and bulk materials. Nanoclusters are a completely new type of building block for thin film deposition. As preformed entities, clusters provide a new means of tailoring the properties of thin films before their growth, simply by changing the size or composition of the clusters that are to be deposited. Contrary to contemporary methods of thin film growth, which mainly rely on the deposition of single atoms, cluster deposition also allows for a more precise assembly of thin films, as the configuration of single atoms with respect to each other is already predetermined in clusters. Nanocluster deposition offers a possibility for the coating of virtually any material with a nanostructured thin film, and therein the enhancement of already existing physical or chemical properties, or the addition of some exciting new feature. A clearer understanding of cluster-surface interactions, and the growth of thin films by cluster deposition, must, however, be achieved, if clusters are to be successfully used in thin film technologies. Using a combination of experimental techniques and molecular dynamics simulations, both the deposition of nanoclusters, and the growth and modification of cluster-assembled thin films, are studied in this thesis. Emphasis is laid on an understanding of the interaction between metal clusters and surfaces, and therein the behaviour of these clusters during deposition and thin film growth. The behaviour of single metal clusters, as they impact on clean metal surfaces, is analysed in detail, from which it is shown that there exists a cluster size and deposition energy dependent limit, below which epitaxial alignment occurs. If larger clusters are deposited at low energies, or cluster-surface interactions are weaker, non-epitaxial deposition will take place, resulting in the formation of nanocrystalline structures. The effect of cluster size and deposition energy on the morphology of cluster-assembled thin films is also determined, from which it is shown that nanocrystalline cluster-assembled films will be porous. Modification of these thin films, with the purpose of enhancing their mechanical properties and durability, without destroying their nanostructure, is presented. Irradiation with heavy ions is introduced as a feasible method for increasing the density, and therein the mechanical stability, of cluster-assembled thin films, without critically destroying their nanocrystalline properties. The results of this thesis demonstrate that nanocluster deposition is a suitable technique for the growth of nanostructured thin films. The interactions between nanoclusters and their supporting surfaces must, however, be carefully considered, if a controlled growth of cluster-assembled thin films, with precisely tailored properties, is to be achieved.

Molecular/Cluster Statistical Thermodynamics Methods To Simulate Quasi-Static Deformations At Finite Temperature

The rapid evolution of nanotechnology appeals for the understanding of global response of nanoscale systems based on atomic interactions, hence necessitates novel, sophisticated, and physically based approaches to bridge the gaps between various length and time scales. In this paper, we propose a group of statistical thermodynamics methods for the simulations of nanoscale systems under quasi-static loading at finite temperature, that is, molecular statistical thermodynamics (MST) method, cluster statistical thermodynamics (CST) method, and the hybrid molecular/cluster statistical thermodynamics (HMCST) method. These methods, by treating atoms as oscillators and particles simultaneously, as well as clusters, comprise different spatial and temporal scales in a unified framework. One appealing feature of these methods is their "seamlessness" or consistency in the same underlying atomistic model in all regions consisting of atoms and clusters, and hence can avoid the ghost force in the simulation. On the other hand, compared with conventional MD simulations, their high computational efficiency appears very attractive, as manifested by the simulations of uniaxial compression and nanoindenation. (C) 2008 Elsevier Ltd. All rights reserved.

Spin-cluster effect and lattice-deformation-induced Kondo effect, spin-glass freezing, and strong phonon scattering in La 0.7 Ca 0.3Mn 1-xCr xO 3

Besides the Kondo effect observed in dilute magnetic alloys, the Cr-doped perovskite manganate compounds La0.7 Ca0.3 Mn1-x Crx O3 also exhibit Kondo effect and spin-glass freezing in a certain composition range. An extensive investigation for the La0.7 Ca0.3 Mn1-x Crx O3 (x=0.01, 0.05, 0.10, 0.3, 0.6, and 1.0) system on the magnetization and ac susceptibility, the resistivity and magnetoresistance, as well as the thermal conductivity is done at low temperature. The spin-glass behavior has been confirmed for these compounds with x=0.05, 0.1, and 0.3. For temperatures above Tf (the spin-glass freezing temperature) a Curie-Weiss law is obeyed. The paramagnetic Curie temperature θ is dependent on Cr doping. Below Tf there exists a Kondo minimum in the resistivity. Colossal magnetoresistance has been observed in this system with Cr concentration up to x=0.6. We suppose that the substitution of Mn with Cr dilutes Mn ions and changes the long-range ferromagnetic order of La0.7 Ca0.3 MnO3. These behaviors demonstrate that short-range ferromagnetic correlation and fluctuation exist among Mn spins far above Tf. Furthermore, these interactions are a precursor of the cooperative freezing at Tf. The "double bumps" feature in the resistivity-temperature curve is observed in compounds with x=0.05 and 0.1. The phonon scattering is enhanced at low temperatures, where the second peak of double bumps comes out. The results indicate that the spin-cluster effect and lattice deformation induce Kondo effect, spin-glass freezing, and strong phonon scattering in mixed perovskite La0.7 Ca0.3 Mn1-x Crx O3. © 2005 American Institute of Physics.

Generation of cluster states in optomechanical quantum systems

We consider an optomechanical quantum system composed of a single cavity mode interacting with N mechanical resonators. We propose a scheme for generating continuous-variable graph states of arbitrary size and shape, including the so-called cluster states for universal quantum computation. The main feature of this scheme is that, differently from previous approaches, the graph states are hosted in the mechanical degrees of freedom rather than in the radiative ones. Specifically, via a 2N-tone drive, we engineer a linear Hamiltonian which is instrumental to dissipatively drive the system to the desired target state. The robustness of this scheme is assessed against finite interaction times and mechanical noise, confirming it as a valuable approach towards quantum state engineering for continuous-variable computation in a solid-state platform.

The cluster initiative greenbook

Competitividad y valor compartido

La production de connaissances dans le cluster montréalais de l’aérospatiale. Le cas d’un réseau local d’innovation

L’importance que revêt la localisation des entreprises dans le fonctionnement et la compétitivité de celles-ci amène à porter l’attention sur la problématique des clusters et l’influence qu’ils peuvent avoir sur la production de connaissances. À travers une perspective théorique qui repose sur la prise en compte des travaux portant sur les districts industriels et ceux plus récents, sur les clusters, nous mettons en évidence une évolution conceptuelle du phénomène de la localisation des entreprises. La place qu’occupe la production de connaissances constitue désormais un élément central des travaux récents qui portent sur les districts et les clusters industriels. Notre examen des dynamiques de fonctionnement de ces systèmes permet d’affirmer que la production de connaissances est une caractéristique centrale des clusters ainsi que leur principal avantage compétitif. Étroitement liée aux réseaux inter-organisationnels, la production de connaissances n’est pas un phénomène naturel, elle découle des mécanismes intrinsèques aux clusters, qu’on cherche à mettre en évidence. Pour ce faire, notre méthodologie qui emprunte les principaux repères de l’analyse stratégique des organisations conduit à l’étude du fonctionnement concret d’un réseau local d’innovation, constitué d’entreprises et d’institutions locales présentes dans le cluster montréalais de l’aérospatiale. Un réseau constitué par l’intermédiaire du Consortium de Recherche et d’Innovation en Aérospatiale du Québec, une institution centrale dans le fonctionnement de ce cluster.

Assessing the Precision of Compositional Data in a Stratified Double Stage Cluster Sample: Application to the Swiss Earnings Structure Survey

Precision of released figures is not only an important quality feature of official statistics, it is also essential for a good understanding of the data. In this paper we show a case study of how precision could be conveyed if the multivariate nature of data has to be taken into account. In the official release of the Swiss earnings structure survey, the total salary is broken down into several wage components. We follow Aitchison's approach for the analysis of compositional data, which is based on logratios of components. We first present diferent multivariate analyses of the compositional data whereby the wage components are broken down by economic activity classes. Then we propose a number of ways to assess precision

Determining feature relevance for the grouping of motor unit action potentials through generative topographic mapping

The study of motor unit action potential (MUAP) activity from electrornyographic signals is an important stage on neurological investigations that aim to understand the state of the neuromuscular system. In this context, the identification and clustering of MUAPs that exhibit common characteristics, and the assessment of which data features are most relevant for the definition of such cluster structure are central issues. In this paper, we propose the application of an unsupervised Feature Relevance Determination (FRD) method to the analysis of experimental MUAPs obtained from healthy human subjects. In contrast to approaches that require the knowledge of a priori information from the data, this FRD method is embedded on a constrained mixture model, known as Generative Topographic Mapping, which simultaneously performs clustering and visualization of MUAPs. The experimental results of the analysis of a data set consisting of MUAPs measured from the surface of the First Dorsal Interosseous, a hand muscle, indicate that the MUAP features corresponding to the hyperpolarization period in the physisiological process of generation of muscle fibre action potentials are consistently estimated as the most relevant and, therefore, as those that should be paid preferential attention for the interpretation of the MUAP groupings.

Metal cluster expansion by addition of low valent group 14 reagents: III. Reaction of bis[bis(trimethylsilyl)methyl]tin with M3(CO)12 or its derivatives (M = Fe, Ru, or Os): the crystal and molecular structures of M3(μ-SnR2)2(CO)10(R = CH(SiMe3)2; M = Ru or Os)

The stannylene [SnR2] (R = CH(SiMe3)2) reacts in different ways with the three dodecacarbonyls of the iron triad: [Fe3(CO)12] gives [Fe2(CO)8(μ-SnR2)], [Ru3(CO)12] gives the planar pentametallic cluster [Ru3(CO)10(μ-SnR2)2], for which a full structural analysis is reported, while [Os3(CO)12] fails to react. Different products are also obtained from three nitrile derivatives: [Fe3-(CO)11(MeCN)] gives [Fe2(CO)6(μ-SnR2)2], which has a structure significantly different from that of known Fe2Sn2 clusters, [Ru3(CO)10(MeCN)2] gives the pentametallic cluster described above, while [Os3(CO)10(MeCN)2] gives the isostructural osmium analogue, which shows the unusual feature of a CO group bridging two osmium atoms.