Kernels semàntics per a clustering de patrons

Memòria elaborada a partir d’una estada al projecte Proteus de la New York University entre abril i juny del 2007. Les tècniques de clustering poden ajudar a reduir la supervisió en processos d’obtenció de patrons per a Extracció d’Informació. Tanmateix, és necessari disposar d’algorismes adequats a documents, i aquests algorismes requereixen mesures adequades de similitud entre patrons. Els kernels poden oferir una solució a aquests problemes, però l’aprenentatge no supervisat requereix d’estrat`egies m´es astutes que l’aprenentatge supervisat per a incorporar major quantitat d’informació. En aquesta memòria, fruit de la meva estada de mes d’Abril al de Juny de 2007 al projecte. Proteus de la New York University, es proposen i avaluen diversos kernels sobre patrons. Ini- cialment s’estudien kernels amb una família de patrons restringits, i a continuació s’apliquen kernels ja usats en tasques supervisades d’Extracció d’Informació. Degut a la degradació del rendiment que experimenta el clustering a l’afegir informació irrellevant, els kernels se simpli- fiquen i es busquen estratègies per a incorporar-hi semàntica de forma selectiva. Finalment, s’estudia quin efecte té aplicar clustering sobre el coneixement semàntic com a pas previ al clustering de patrons. Les diverses estratègies s’avaluen en tasques de clustering de documents i patrons usant dades reals.

Stochastic Search Variable Selection in Vector Error Correction Models with an Application to a Model of the UK Macroeconomy

This paper develops methods for Stochastic Search Variable Selection (currently popular with regression and Vector Autoregressive models) for Vector Error Correction models where there are many possible restrictions on the cointegration space. We show how this allows the researcher to begin with a single unrestricted model and either do model selection or model averaging in an automatic and computationally efficient manner. We apply our methods to a large UK macroeconomic model.

Codex Search: propuesta e implementación de un sistema de recuperación de información de ámbito local basado en casos

CODEX SEARCH es un motor de recuperación de información especializado en derecho de extranjería que está basado en herramientas y conocimiento lingüísticos. Un motor o Sistema de Recuperación de Información (SRI) es un software capaz de localizar información en grandes colecciones documentales (entorno no trivial) en formato electrónico. Mediante un estudio previo se ha detectado que la extranjería es un ámbito discursivo en el que resulta difícil expresar la necesidad de información en términos de una consulta formal, objeto de los sistemas de recuperación actuales. Por lo tanto, para desarrollar un SRI eficiente en el dominio indicado no basta con emplear un modelo tradicional de RI, es decir, comparar los términos de la pregunta con los de la respuesta, básicamente porque no expresan implicaciones y porque no tiene que haber necesariamente una relación 1 a 1. En este sentido, la solución lingüística propuesta se basa en incorporar el conocimiento del especialista mediante la integración en el sistema de una librería de casos. Los casos son ejemplos de procedimientos aplicados por expertos a la solución de problemas que han ocurrido en la realidad y que han terminado en éxito o fracaso. Los resultados obtenidos en esta primera fase son muy alentadores pero es necesario continuar la investigación en este campo para mejorar el rendimiento del prototipo al que se puede acceder desde &http://161.116.36.139/~codex/&.

QWERTY and the search for optimality

This paper shows how one of the developers of QWERTY continued to use the trade secret that underlay its development to seek further efficiency improvements after its introduction. It provides further evidence that this was the principle used to design QWERTY in the first place and adds further weight to arguments that QWERTY itself was a consequence of creative design and an integral part of a highly efficient system rather than an accident of history. This further serves to raise questions over QWERTY's forced servitude as 'paradigm case' of inferior standard in the path dependence literature. The paper also shows how complementarities in forms of intellectual property rights protection played integral roles in the development of QWERTY and the search for improvements on it, and also helped effectively conceal the source of the efficiency advantages that QWERTY helped deliver.

Efficient Competition through Cheap Talk: Competing Auctions and Competitive Search without Ex Ante Price

We consider a frictional two-sided matching market in which one side uses public cheap talk announcements so as to attract the other side. We show that if the first-price auction is adopted as the trading protocol, then cheap talk can be perfectly informative, and the resulting market outcome is efficient, constrained only by search frictions. We also show that the performance of an alternative trading protocol in the cheap-talk environment depends on the level of price dispersion generated by the protocol: If a trading protocol compresses (spreads) the distribution of prices relative to the first-price auction, then an efficient fully revealing equilibrium always (never) exists. Our results identify the settings in which cheap talk can serve as an efficient competitive instrument, in the sense that the central insights from the literature on competing auctions and competitive search continue to hold unaltered even without ex ante price commitment.

Large Bayesian VARMAs

Vector Autoregressive Moving Average (VARMA) models have many theoretical properties which should make them popular among empirical macroeconomists. However, they are rarely used in practice due to over-parameterization concerns, difficulties in ensuring identification and computational challenges. With the growing interest in multivariate time series models of high dimension, these problems with VARMAs become even more acute, accounting for the dominance of VARs in this field. In this paper, we develop a Bayesian approach for inference in VARMAs which surmounts these problems. It jointly ensures identification and parsimony in the context of an efficient Markov chain Monte Carlo (MCMC) algorithm. We use this approach in a macroeconomic application involving up to twelve dependent variables. We find our algorithm to work successfully and provide insights beyond those provided by VARs.

Meeting technologies and optimal trading mechanisms in competitive search markets

In a market in which sellers compete by posting mechanisms, we study how the properties of the meeting technology affect the mechanism that sellers select. In general, sellers have incentive to use mechanisms that are socially efficient. In our environment, sellers achieve this by posting an auction with a reserve price equal to their own valuation, along with a transfer that is paid by (or to) all buyers with whom the seller meets. However, we define a novel condition on meeting technologies, which we call “invariance,” and show that the transfer is equal to zero if and only if the meeting technology satisfies this condition.

Directed Search over the Life Cycle

We develop a life-cycle model of the labor market in which different worker-firm matches have different quality and the assignment of the right workers to the right firms is time consuming because of search and learning frictions. The rate at which workers move between unemployment, employment and across different firms is endogenous because search is directed and, hence, workers can choose whether to seek low-wage jobs that are easy to find or high-wage jobs that are hard to find. We calibrate our theory using data on labor market transitions aggregated across workers of different ages. We validate our theory by showing that it predicts quite well the pattern of labor market transitions for workers of different ages. Finally, we use our theory to decompose the age profiles of transition rates, wages and productivity into the effects of age variation in work-life expectancy, human capital and match quality.

Fiscal multipliers in a two-sector search and matching model

This paper evaluates the effects of policy interventions on sectoral labour markets and the aggregate economy in a business cycle model with search and matching frictions. We extend the canonical model by including capital-skill complementarity in production, labour markets with skilled and unskilled workers and on-the-job-learning (OJL) within and across skill types. We first find that, the model does a good job at matching the cyclical properties of sectoral employment and the wage-skill premium. We next find that vacancy subsidies for skilled and unskilled jobs lead to output multipliers which are greater than unity with OJL and less than unity without OJL. In contrast, the positive output effects from cutting skilled and unskilled income taxes are close to zero. Finally, we find that the sectoral and aggregate effects of vacancy subsidies do not depend on whether they are financed via public debt or distorting taxes.

The antecedents and innovation consequences of organizational search: empirical evidence for Spain

This paper examines the antecedents and innovation consequences of the methods firms adopt in organizing their search strategies. From a theoretical perspective, organizational search is described using a typology that shows how firms implement exploration and exploitation search activities that span their organizational boundaries. This typology includes three models of implementation: ambidextrous, specialized, and diversified implementation. From an empirical perspective, the paper examines the performance consequences when applying these models, and compares their capacity to produce complementarities. Additionally, since firms' choices in matters of organizational search are viewed as endogenous variables, the paper examines the drivers affecting them and identifies the importance of firms' absorptive capacity and diversified technological opportunities in determining these choices. The empirical design of the paper draws on new data for manufacturing firms in Spain, surveyed between 2003 and 2006.

The rational search for natural neolignans

A rational method of search for natural neolignans of desired structures is outlined. This involves consultation of a collection of chemical profiles of plant families. The profiles are assembled considering the biosynthetic class (in the present case lignoids), subclass (neolignans), structural types (neolignan skeleta) and relative frequency of substitutional derivatives belonging to each type (known compounds). The method is of course applicable to ani class of natural products. Its use in the case of neolignans is here selected as an exemple in view of the recently discovered antagonism towards PAF of kadsurenone, a representative of this subclass of phytochemicals. Application of the chemical profiles to phylogenetic studies is illustrated.

Alineamiento múltiple de secuencias con T-Coffee: una aproximación paralela

Las aplicaciones de alineamiento múltiple de secuencias son prototipos de aplicaciones que requieren elevada potencia de cómputo y memoria. Se destacan por la relevancia científica que tienen los resultados que brindan a investigaciones científicas en el campo de la biomedicina, genética y farmacología. Las aplicaciones de alineamiento múltiple tienen la limitante de que no son capaces de procesar miles de secuencias, por lo que se hace necesario crear un modelo para resolver la problemática. Analizando el volumen de datos que se manipulan en el área de las ciencias biológica y la complejidad de los algoritmos de alineamiento de secuencias, la única vía de solución del problema es a través de la utilización de entornos de cómputo paralelos y la computación de altas prestaciones. La investigación realizada por nosotros tiene como objetivo la creación de un modelo paralelo que le permita a los algoritmos de alineamiento múltiple aumentar el número de secuencias a procesar, tratando de mantener la calidad en los resultados para garantizar la precisión científica. El modelo que proponemos emplea como base la clusterización de las secuencias de entrada utilizando criterios biológicos que permiten mantener la calidad de los resultados. Además, el modelo se enfoca en la disminución del tiempo de cómputo y consumo de memoria. Para presentar y validar el modelo utilizamos T-Coffee, como plataforma de desarrollo e investigación. El modelo propuesto pudiera ser aplicado a cualquier otro algoritmo de alineamiento múltiple de secuencias.

An objective anisotropic elastic plastic model and Algorithm applicable to bone mechanics

The implicit projection algorithm of isotropic plasticity is extended to an objective anisotropic elastic perfectly plastic model. The recursion formula developed to project the trial stress on the yield surface, is applicable to any non linear elastic law and any plastic yield function.A curvilinear transverse isotropic model based on a quadratic elastic potential and on Hill's quadratic yield criterion is then developed and implemented in a computer program for bone mechanics perspectives. The paper concludes with a numerical study of a schematic bone-prosthesis system to illustrate the potential of the model.

Chemical and pharmacological investigation of Solanum species of Brazil: a search for solasodine and other potentially useful therapeutic agents

A systematic search for solasodine, an important staring material for the partial synthesis of steroidal hormones as well as other potentially bioactive constituents of various Solanum species of Brazil has been undertaken. Thus, the fruits of S. paludosum, S. asperum, S. sessiliforum and Solanum sp. were found to contain significant amounts of solasodine. The root bark of S. paludosum which showe durare like activity yelded tomatidenol and another yet unidentified alkaloid responsible for the biological activity. The fruits of S. asperum yelded a new spirosolane alkaloid, solaparnaine. The stem bark of S. pseudo-quina showed convulsive and exitatory activity from which (25S)-isosolafloridine was identified as the active principle. In addition, the latter alkaloid was also found to show antimicrobial activity.

Polyanalgesic Consensus Conference--2012: recommendations on trialing for intrathecal (intraspinal) drug delivery: report of an interdisciplinary expert panel.

INTRODUCTION: Trialing for intrathecal pump placement is an essential part of the decision-making process in placing a permanent device. In both the United States and the international community, the proper method for trialing is ill defined. METHODS: The Polyanalgesic Consensus Conference (PACC) is a group of well-published experienced practitioners who meet to update the state of care for intrathecal therapies on the basis of current knowledge in the literature and clinical experience. Anexhaustive search is performed to create a base of information that the panel considers when making recommendations for best clinical practices. This literature, coupled with clinical experience, is the basis for recommendations and for identification of gaps in the base of knowledge regarding trialing for intrathecal pump placement. RESULTS: The panel has made recommendations for the proper methods of trialing for long-term intrathecal drug delivery. CONCLUSION: The use of intrathecal drug delivery is an important part of the treatment algorithm for moderate to severe chronic pain. It has become common practice to perform a temporary neuroaxial infusion before permanent device implantation. On the basis of current knowledge, the PACC has developed recommendations to improve care. The need to update these recommendations will be very important as new literature is published.