MULTI-INSTRUMENT DETECTION IN POLYPHONIC MUSIC USING GAUSSIAN MIXTURE BASED FACTORIAL HMM

We formulate the problem of detecting the constituent instruments in a polyphonic music piece as a joint decoding problem. From monophonic data, parametric Gaussian Mixture Hidden Markov Models (GM-HMM) are obtained for each instrument. We propose a method to use the above models in a factorial framework, termed as Factorial GM-HMM (F-GM-HMM). The states are jointly inferred to explain the evolution of each instrument in the mixture observation sequence. The dependencies are decoupled using variational inference technique. We show that the joint time evolution of all instruments' states can be captured using F-GM-HMM. We compare performance of proposed method with that of Student's-t mixture model (tMM) and GM-HMM in an existing latent variable framework. Experiments on two to five polyphony with 8 instrument models trained on the RWC dataset, tested on RWC and TRIOS datasets show that F-GM-HMM gives an advantage over the other considered models in segments containing co-occurring instruments.

Dynamical generation of flavour

We propose the generation of Standard Model fermion hierarchy by the extension of renormalizable SO(10) GUT with O(N (g) ) family gauge symmetry. In this scenario, Higgs representations of SO(10) also carry family indices and are called Yukawons. Vacuum expectation values of these Yukawon fields break GUT and family symmetry and generate MSSM Yukawa couplings dynamically. We have demonstrated this idea using Higgs irrep, ignoring the contribution of 1 2 0-plet which is, however, required for complete fitting of fermion mass-mixing data. The effective MSSM matter fermion couplings to the light Higgs pair are determined by the null eigenvectors of the MSSM-type Higgs doublet superfield mass matrix . A consistency condition on the doublet (1,2,+/- 1]) mass matrix ( 0) is required to keep one pair of Higgs doublets light in the effective MSSM. We show that the Yukawa structure generated by null eigenvectors of are of generic kind required by the MSSM. A hidden sector with a pair of (S (a b) ; I center dot (a b) ) fields breaks supersymmetry and facilitates 0. SUSY breaking is communicated via supergravity. In this scenario, matter fermion Yukawa couplings are reduced from 15 to just 3 parameters in MSGUT with three generations.

Gravitational rescue of minimal gauge mediation

Gravity mediated supersymmetry breaking becomes comparable to gauge mediated supersymmetry breaking contributions when messenger masses are close to the GUT scale. By suitably arranging the gravity contributions, one can modify the soft supersymmetry breaking sector to generate a large stop mixing parameter and a light Higgs mass of 125 GeV. In this kind of hybrid models, however, the nice features of gauge mediation like flavor conservation, etc. are lost. To preserve the nice features, gravitational contributions should become important for lighter messenger masses and should be important only for certain fields. This is possible when the hidden sector contains multiple (at least two) spurions with hierarchical vacuum expectation values. In this case, the gravitational contributions can be organized to be ``just right.'' We present a complete model with two spurion hidden sector where the gravitational contribution is from a warped flavor model in a Randall-Sundrum setting. Along the way, we present simple expressions to handle renormalization group equations when supersymmetry is broken by two different sectors at two different scales.

Simulation-based optimal sensor scheduling with application to observer trajectory planning

The sensor scheduling problem can be formulated as a controlled hidden Markov model and this paper solves the problem when the state, observation and action spaces are continuous. This general case is important as it is the natural framework for many applications. The aim is to minimise the variance of the estimation error of the hidden state w.r.t. the action sequence. We present a novel simulation-based method that uses a stochastic gradient algorithm to find optimal actions. © 2007 Elsevier Ltd. All rights reserved.

La sabiduría desde el “Quaerens intellectum” en Anselmo de Canterbury

Resumen: La prueba de la existencia de Dios, sacada de su sola idea, fue, y es aún, una de esas experiencias metafísicas de las que se puede decir que nacen eternas, porque alcanzan el último término de uno de los caminos por donde el espíritu humano puede encauzarse. La presencia de Dios en el alma vivifica la fórmula que da san Anselmo y el análisis que hace de ella no tiene otro fin que llevar a la clara luz de la inteligencia una presencia realmente, pero oscuramente, sentida. Situados, dice, entre la fe y la visión, considera que “la inteligencia que alcanzamos en esta vida es intermedia”; los modos de aproximación difieren, pero sus términos no.

La heráldica, el retrato y el fin de la Edad Media : un acercamiento antropológico a la iconografía

Resumen: Durante la Edad Media hizo su aparición la heráldica como forma de identificar ciertos aspectos de las personas, ya sea por su condición social o por su actividad específica. El uso de blasones y escudos de armas desde el siglo XI generó una importante cantidad de signos, colores y símbolos que edificaron todo un “patrón de lectura” del “rostro social” de la persona que portaba tal insignia, sin importar o no la veracidad de su rostro al confeccionarse un retrato pintado o esculpido. Paralelo a esto, la retratística que hubo durante todo el período medieval de personas reales, sagradas o ideales avanzó hacia un progresivo naturalismo que finalizó en el siglo XV con la reproducción del retratado tal cual era su fisonomía y abandonando el canon estereotipado que durante los siglos precedentes habían proliferado. Finalmente, con el creciente Humanismo, la aparición de la perspectiva y el progresivo abandono de la identificación de determinado personaje con sólo su escudo de armas, sirvió para que el retratado mirase a los ojos del observador y “dialogase”, develando aspectos espirituales antes ocultos. El objetivo de este trabajo es acercarse con una visión antropológica a la cuestión del retrato tardomedieval y renacentista en conjunción con los cambios operados en la heráldica y la iconografía, comprobando que la mirada tiene marcada intencionalidad, los rostros se frontalizan progresivamente y la heráldica pierde el espesor que tuvo en siglos precedentes

Asymptotic behaviour of approximate Bayesian estimators

Although approximate Bayesian computation (ABC) has become a popular technique for performing parameter estimation when the likelihood functions are analytically intractable there has not as yet been a complete investigation of the theoretical properties of the resulting estimators. In this paper we give a theoretical analysis of the asymptotic properties of ABC based parameter estimators for hidden Markov models and show that ABC based estimators satisfy asymptotically biased versions of the standard results in the statistical literature.

Filtering via approximate Bayesian computation

Approximate Bayesian computation (ABC) has become a popular technique to facilitate Bayesian inference from complex models. In this article we present an ABC approximation designed to perform biased filtering for a Hidden Markov Model when the likelihood function is intractable. We use a sequential Monte Carlo (SMC) algorithm to both fit and sample from our ABC approximation of the target probability density. This approach is shown to, empirically, be more accurate w.r.t.~the original filter than competing methods. The theoretical bias of our method is investigated; it is shown that the bias goes to zero at the expense of increased computational effort. Our approach is illustrated on a constrained sequential lasso for portfolio allocation to 15 constituents of the FTSE 100 share index.

La apariencia como trans-parencia : una situación existencial del homo poeticus

Resumen: Dostoevskij atribuye a la belleza una función salvífica (“la belleza salvará al mundo”), poniéndola así en relación con el campo afectivo, salvación significa en efecto vida plena; para Matisse “Todo arte digno de tal nombre es religioso”, es decir, posee una función de “mediación-unión” entre órdenes diversos. La noción de “visión interior”, es la que mejor fundamenta estas convicciones. Ella atraviesa el tiempo y el espacio: está presente en la antigua cultura china, así como en la Grecia clásica y en la Antigüedad Tardía y permanece implícita durante todo el Medioevo; se la vuelve a encontrar en alguno de los máximos pintores modernos (Kandinsky, Chagal). Todos los testimonios de los artistas concuerdan en indicar que para alcanzarla es necesario una especie de ascesis: ésta consiste en una revelación, experimentada como un don, del Sentido oculto en lo sensible; por el hecho de dar inicio al pasaje de un nivel ontológico a otro (valencia re-ligiosa o syn-bolica) y de consistir en una experiencia de plenitud de vida, le pertenece además un efecto transformante (valencia afectiva). Esta vía propiamente humana de la “visión interior”, por la cual la apariencia (lo sensible) se vuelve trans-parencia de lo invisible, se presenta como la más adecuada al arte cristiano para el culto.

Translatio materiae : adopción de formas y géneros latinos en las colecciones hispánicas de milagros marianos

Resumen: Un tipo de literatura latina no narrativa, asimilable más bien al discurso doctrinal teológico y mariológico y, por otro lado, a la himnodia litúrgica, deja entrever su influencia en las colecciones ibéricas de milagros marianos en verso del siglo XIII, las Cantigas de Santa María de Alfonso X y los Milagros de Nuestra Señora de Gonzalo de Berceo. Las muy escasas menciones a fuentes en las Cantigas de Santa María no invalidan la evidencia de múltiples “citas” reconocibles. Así, por ejemplo, es posible encontrar himnos y antífonas explícitamente citados (Te Deum laudamus, Salve Regina, Ave Maris Stella, además de glosas al Ave María) y, más allá de estos, se perciben otros modos de inserción de la lírica himnódica en la narración de los milagros. Se propone, entonces, iluminar estos loci donde parece haber un texto aludido o “escondido”, para considerar sus diversos modos de traducción, adopción y reelaboración.

Distributed maximum likelihood for self-localization in sensor networks

We show that the sensor localization problem can be cast as a static parameter estimation problem for Hidden Markov Models and we develop fully decentralized versions of the Recursive Maximum Likelihood and the Expectation-Maximization algorithms to localize the network. For linear Gaussian models, our algorithms can be implemented exactly using a distributed version of the Kalman filter and a message passing algorithm to propagate the derivatives of the likelihood. In the non-linear case, a solution based on local linearization in the spirit of the Extended Kalman Filter is proposed. In numerical examples we show that the developed algorithms are able to learn the localization parameters well.

Sequential decision making in general state space models

Many problems in control and signal processing can be formulated as sequential decision problems for general state space models. However, except for some simple models one cannot obtain analytical solutions and has to resort to approximation. In this thesis, we have investigated problems where Sequential Monte Carlo (SMC) methods can be combined with a gradient based search to provide solutions to online optimisation problems. We summarise the main contributions of the thesis as follows. Chapter 4 focuses on solving the sensor scheduling problem when cast as a controlled Hidden Markov Model. We consider the case in which the state, observation and action spaces are continuous. This general case is important as it is the natural framework for many applications. In sensor scheduling, our aim is to minimise the variance of the estimation error of the hidden state with respect to the action sequence. We present a novel SMC method that uses a stochastic gradient algorithm to find optimal actions. This is in contrast to existing works in the literature that only solve approximations to the original problem. In Chapter 5 we presented how an SMC can be used to solve a risk sensitive control problem. We adopt the use of the Feynman-Kac representation of a controlled Markov chain flow and exploit the properties of the logarithmic Lyapunov exponent, which lead to a policy gradient solution for the parameterised problem. The resulting SMC algorithm follows a similar structure with the Recursive Maximum Likelihood(RML) algorithm for online parameter estimation. In Chapters 6, 7 and 8, dynamic Graphical models were combined with with state space models for the purpose of online decentralised inference. We have concentrated more on the distributed parameter estimation problem using two Maximum Likelihood techniques, namely Recursive Maximum Likelihood (RML) and Expectation Maximization (EM). The resulting algorithms can be interpreted as an extension of the Belief Propagation (BP) algorithm to compute likelihood gradients. In order to design an SMC algorithm, in Chapter 8 uses a nonparametric approximations for Belief Propagation. The algorithms were successfully applied to solve the sensor localisation problem for sensor networks of small and medium size.

An approximate likelihood method for estimating static parameters in multi-target tracking models

This paper considers a class of dynamic Spatial Point Processes (PP) that evolves over time in a Markovian fashion. This Markov in time PP is hidden and observed indirectly through another PP via thinning, displacement and noise. This statistical model is important for Multi object Tracking applications and we present an approximate likelihood based method for estimating the model parameters. The work is supported by an extensive numerical study.

Cidadãos crentes, críticos e ocultos : assimetrias da confiança no Congresso Nacional

A literatura atribui a confiança e a desconfiança no Congresso Nacional ao resultado da avaliação do desempenho dos parlamentares. Pondera que, desde que associada ao crescimento da escolarização, da renda e da adesão à democracia, a desconfiança nas instituições democráticas pode refletir maior exigência de cidadãos críticos, ou democratas insatisfeitos, advindos de melhoria social. Nesta perspectiva, a reprovação é positiva, contanto que fenômeno específico, e potencialmente propulsora do engajamento cívico dos cidadãos, fundamental para a democracia nas chaves representativa e participativa. Na busca de uma cidadania crítica à brasileira, a pesquisa que origina o artigo analisou as segmentações de escolaridade e de renda da avaliação de desempenho parlamentar do Datafolha entre 2005 e 2008, período em que os cidadãos tiveram grande oferta de informação (no caso, negativa)sobre o Congresso Nacional e que concentrou dois dos três piores índices de reprovação da história da pesquisa, associados a eventos do escândalo do mensalão. Acabou por evidenciar que, além dos cidadãos críticos, nossas desigualdades parecem ter forjado outros dois tipos de cidadania, muitas vezes relevados pela literatura: a crente e a oculta, potencialmente prejudiciais ao regime, pela associação à baixa escolaridade e à baixa renda, pelo reforço à exclusão e pela própria invisibilidade. Preteridos na academia, não podem sê-lo pela "Casa de todos os brasileiros" em um País que vislumbra alcançar indicadores sociais próximos às democracias centrais na próxima década.

Sociedades Cooperativas. Aspectos Contables Singulares

[ES] Las sociedades cooperativas han visto recientemente reforzada su normativa contable con la aprobación de la ORDEN ECO/3614/2003 que tiene por objeto la adaptación de las Normas de valoración y elaboración de cuentas anuales aprobada en el Plan General de Contabilidad a las características de estas entidades.