Fast cross-validation of kernel fisher discriminant classifiers

Given n training examples, the training of a Kernel Fisher Discriminant (KFD) classifier corresponds to solving a linear system of dimension n. In cross-validating KFD, the training examples are split into 2 distinct subsets for a number of times (L) wherein a subset of m examples is used for validation and the other subset of(n - m) examples is used for training the classifier. In this case L linear systems of dimension (n - m) need to be solved. We propose a novel method for cross-validation of KFD in which instead of solving L linear systems of dimension (n - m), we compute the inverse of an n × n matrix and solve L linear systems of dimension 2m, thereby reducing the complexity when L is large and/or m is small. For typical 10-fold and leave-one-out cross-validations, the proposed algorithm is approximately 4 and (^4/9n) times respectively as efficient as the naive implementations. Simulations are provided to demonstrate the efficiency of the proposed algorithms.

Graph matching : fast candidate elimination using machine learning techniques

Graph matching is an important class of methods in pattern recognition. Typically, a graph representing an unknown pattern is matched with a database of models. If the database of model graphs is large, an additional factor in induced into the overall complexity of the matching process. Various techniques for reducing the influence of this additional factor have been described in the literature. In this paper we propose to extract simple features from a graph and use them to eliminate candidate graphs from the database. The most powerful set of features and a decision tree useful for candidate elimination are found by means of the C4.5 algorithm, which was originally proposed for inductive learning of classication rules. Experimental results are reported demonstrating that effcient candidate elimination can be achieved by the proposed procedure.

International collaborative project to compare and track the nutritional composition of fast foods

Background
Chronic diseases are the leading cause of premature death and disability in the world with over-nutrition a primary cause of diet-related ill health. Excess quantities of energy, saturated fat, sugar and salt derived from fast foods contribute importantly to this disease burden. Our objective is to collate and compare nutrient composition data for fast foods as a means of supporting improvements in product formulation.
Methods/design
Surveys of fast foods will be done in each participating country each year. Information on the nutrient composition for each product will be sought either through direct chemical analysis, from fast food companies, in-store materials or from company websites. Foods will be categorized into major groups for the primary analyses which will compare mean levels of saturated fat, sugar, sodium, energy and serving size at baseline and over time. Countries currently involved include Australia, New Zealand, France, UK, USA, India, Spain, China and Canada, with more anticipated to follow.
Discussion
This collaborative approach to the collation and sharing of data will enable low-cost tracking of fast food composition around the world. This project represents a significant step forward in the objective and transparent monitoring of industry and government commitments to improve the quality of fast foods.

Fabrication of polycarbonate microfluidic devices using a focusing and defocusing methods of CO2 laser system

Fabrication of microfluidic devices has always been a challenging endeavour due to its characteristics and cost involved. Recently, laser ablation using low cost engravers has been exploited for fast and cheap prototyping. In this paper, we explore the various parameters affecting the fabrication of polycarbonate microfluidic channels using CO2 laser ablation. The results show that, by manipulating the focus length of the laser beam, we can achieve good cross sectional profiles with low surface roughness. The results also show that various profiles can be fabricated by changing the laser parameters such as laser power, cutting speed and number of laser pulses.

Very high pressure liquid chromatography using fully porous particles: Quantitative analysis of fast gradient separations without post-run times

Using a column packed with fully porous particles, four methods for controlling the flow rates at which gradient elution runs are conducted in very high pressure liquid chromatography (VHPLC) were tested to determine whether reproducible thermal conditions could be achieved, such that subsequent analyses would proceed at nearly the same initial temperature. In VHPLC high flow rates are achieved, producing fast analyses but requiring high inlet pressures. The combination of high flow rates and high inlet pressures generates local heat, leading to temperature changes in the column. Usually in this case a post-run time is input into the analytical method to allow the return of the column temperature to its initial state. An alternative strategy involves operating the column without a post-run equilibration period and maintaining constant temperature variations for subsequent analysis after conducting one or a few separations to bring the column to a reproducible starting temperature. A liquid chromatography instrument equipped with a pressure controller was used to perform constant pressure and constant flow rate VHPLC separations. Six replicate gradient separations of a nine component mixture consisting of acetophenone, propiophenone, butyrophenone, valerophenone, hexanophenone, heptanophenone, octanophenone, benzophenone, and acetanilide dissolved in water/acetonitrile (65:35, v/v) were performed under various experimental conditions: constant flow rate, two sets of constant pressure, and constant pressure operation with a programmed flow rate. The relative standard deviations of the response factors for all the analytes are lower than 5% across the methods. Programming the flow rate to maintain a fairly constant pressure instead of using instrument controlled constant pressure improves the reproducibility of the retention times by a factor of 5, when plotting the chromatograms in time.

New opportunities for quantitative and time efficient 3D MRI of liquid and solid electrochemical cell components: Sectoral Fast Spin Echo and SPRITE.

The ability to image electrochemical processes in situ using nuclear magnetic resonance imaging (MRI) offers exciting possibilities for understanding and optimizing materials in batteries, fuel cells and supercapacitors. In these applications, however, the quality of the MRI measurement is inherently limited by the presence of conductive elements in the cell or device. To overcome related difficulties, optimal methodologies have to be employed. We show that time-efficient three dimensional (3D) imaging of liquid and solid lithium battery components can be performed by Sectoral Fast Spin Echo and Single Point Imaging with T1 Enhancement (SPRITE), respectively. The former method is based on the generalized phase encoding concept employed in clinical MRI, which we have adapted and optimized for materials science and electrochemistry applications. Hard radio frequency pulses, short echo spacing and centrically ordered sectoral phase encoding ensure accurate and time-efficient full volume imaging. Mapping of density, diffusivity and relaxation time constants in metal-containing liquid electrolytes is demonstrated. 1, 2 and 3D SPRITE approaches show strong potential for rapid high resolution (7)Li MRI of lithium electrode components.

Fast one-class support vector machine for novelty detection

Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.

The art of being together: a social constructionist perspective on dialogic methods in the organizational context

The purpose of this project is to understand, under a social constructionist approach, what are the meanings that external facilitators and organizational members (sponsors) working with dialogic methods place on themselves and their work. Dialogic methods, with the objective of engaging groups in flows of conversations to envisage and co-create their own future, are growing fast within organizations as a means to achieve collective change. Sharing constructionist ideas about the possibility of multiple realities and language as constitutive of such realities, dialogue has turned into a promising way for transformation, especially in a macro context of constant change and increasing complexity, where traditional structures, relationships and forms of work are questioned. Research on the topic has mostly focused on specific methods or applications, with few attempts to study it in a broader sense. Also, despite the fact that dialogic methods work on the assumption that realities are socially constructed, few studies approach the topic from a social constructionist perspective, as a research methodology per se. Thus, while most existing research aims at explaining whether or how particular methods meet particular results, my intention is to explore the meanings sustaining these new forms of organizational practice. Data was collected through semi-structured interviews with 25 people working with dialogic methods: 11 facilitators and 14 sponsors, from 8 different organizations in Brazil. Firstly, the research findings indicate several contextual elements that seem to sustain the choices for dialogic methods. Within this context, there does not seem to be a clear or specific demand for dialogic methods, but a set of different motivations, objectives and focuses, bringing about several contrasts in the way participants name, describe and explain their experiences with such methods, including tensions on power relations, knowledge creation, identity and communication. Secondly, some central ideas or images were identified within such contrasts, pointing at both directions: dialogic methods as opportunities for the creation of new organizational realities (with images of a ‘door’ or a ‘flow’, for instance, which suggest that dialogic methods may open up the access to other perspectives and the creation of new realities); and dialogic methods as new instrumental mechanisms that seem to reproduce the traditional and non-dialogical forms of work and relationship. The individualistic tradition and its tendency for rational schematism - pointed out by social constructionist scholars as strong traditions in our Western Culture - could be observed in some participants’ accounts with the image of dialogic methods as a ‘gym’, for instance, in which dialogical – and idealized –‘abilities’ could be taught and trained, turning dialogue into a tool, rather than a means for transformation. As a conclusion, I discuss what the implications of such taken-for-granted assumptions may be, and offer some insights into dialogue (and dialogic methods) as ‘the art of being together’.

Methodology for fast evaluation of Bacillus thuringiensis crystal protein content

O desenvolvimento da produção e uso do Bacillus thuringiensis no Brasil em escala comercial enfrenta certas dificuldades, entre elas o estabelecimento de metodologias para a quantificação de produtos tóxicos a serem comercializados. Atualmente, a quantidade de toxinas é expressa como porcentagem do total de proteínas presentes em amostras em consideração. Tal metodologia, entretanto, não mede a quantidade real de uma determinada proteína presente em um produto qualquer, além do fato de diferentes linhagens bacterianas possuírem diferentes genes codificadores para endotoxinas e mesmo para b-toxina. Desde que os diferentes tipos de toxinas apresentam diferentes características antigências, este trabalho tem como objetivo a utilização de técnicas imunológicas para quantificar específicamente o conteúdo de proteína cristal presente em diferentes amostras. A proteína cristal produzida pela subespécie B. thuringiensis var. israelensis foi purificada por ultracentrifugação e utilizada para imunizar coelhas e produzir soros hiperimunes. Tais soros foram posteriormente usados para avaliar o nível de proteína cristal em bioinseticidas comerciais e em culturas de laboratório desta bactéria utilizando-se a técnica do imunodot. Os resultados foram obtidos por comparação de reações com concentrações conhecidas de proteína cristal permitindo assim avaliar com segurança os níveis desta proteína em várias preparações.

Fast-forward/fast-backward substitutions on vector computers

This paper deals with approaches for sparse matrix substitutions using vector processing. Many publications have used the W-matrix method to solve the forward/backward substitutions on vector computer. Recently a different approach has been presented using dependency-based substitution algorithm (DBSA). In this paper the focus is on new algorithms able to explore the sparsity of the vectors. The efficiency is tested using linear systems from power systems with 118, 320, 725 and 1729 buses. The tests were performed on a CRAY Y MP2E/232. The speedups for a fast-forward/fast-backward using a 1729-bus system are near 19 and 14 for real and complex arithmetic operations, respectively. When forward/backward is employed the speedups are about 8 and 6 to perform the same simulations.

Continuation fast decoupled power flow with secant predictor

The conventional Newton and fast decoupled power flow methods are considered inadequate for obtaining the maximum loading point of power systems due to ill-conditioning problems at and near this critical point. At this point, the Jacobian matrix of the Newton method becomes singular. In addition, it is widely accepted that the P-V and Q-theta decoupling assumptions made for the fast decoupled power flow formulation no longer hold. However, in this paper, it is presented a new fast decoupled power flow that becomes adequate for the computation of the maximum loading point by simply using the reactive power injection of a selected PV bus as a continuation parameter. Besides, fast decoupled methods using V and 0 as parameters and a secant predictor are also presented. These new versions are compared to each other with the purpose of pointing out their features, as well as the influence of reactive power and transformer tap limits. The results obtained for the IEEE systems (14 and 118 buses) show that the characteristics of the conventional method are enhanced and the region of convergence around the singular solution is enlarged.

Effect of monomer treatment and polymerisation methods on the bond strength of resin teeth to denture base material

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Bond strength of denture teeth to acrylic resin: effect of thermocycling and polymerisation methods

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)