Papers by DAVID MONTALUISA TORRES
A musically meaningful vocabulary is one of the keystones in building a computer audition system ... more A musically meaningful vocabulary is one of the keystones in building a computer audition system that can model the semantics of audio content. If a word in the vocabulary is inconsistently used by human annotators, or the word is not clearly represented by the underlying acoustic representation, the word can be considered as noisy and should be removed from the vocabulary to denoise the modeling process. This paper proposes an approach to construct a vocabulary of predictive semantic concepts based on sparse canonical component analysis (sparse CCA) . Experimental results illustrate that, by identifying musically meaningful words, we can improve the performance of a previously proposed computer audition system for music annotation and retrieval.
Journal of Symplectic Geometry, 2017
A Poisson manifold (M 2n , π) is b-symplectic if n π is transverse to the zero section. In this p... more A Poisson manifold (M 2n , π) is b-symplectic if n π is transverse to the zero section. In this paper we apply techniques native to Symplectic Topology to address questions pertaining to b-symplectic manifolds. We provide constructions of b-symplectic structures on open manifolds by Gromov's h-principle, and of b-symplectic manifolds with a prescribed singular locus, by means of surgeries.
Soldagem & Inspeção, 2016
Resumen La Microscopía de Fuerza Atómica (MFA), técnica de caracterización que genera imágenes to... more Resumen La Microscopía de Fuerza Atómica (MFA), técnica de caracterización que genera imágenes topográficas de superficies a muy altas resoluciones, opera registrando los detalles de relieve superficial del material con un cantiléver que se mueve sobre la muestra, mientras un detector piezoeléctrico monitorea los cambios de altura. Ésta técnica, adquiere relevancia en el campo de la identificación de fases, partículas y precipitados a niveles de nanoescala por lo que se propone para su identificación en la zona de fusión de componentes soldados y para la identificación temprana de fallas por agrietamiento. Es una técnica que no requiere de muestras conductoras, o la especial preparación metalográfica de probetas como es el caso de la microscopia electrónica. El material de estudio es un acero inoxidable dúplex 2205, unido por arco metálico y electrodo de tungsteno (GTAW). Al material soldado se le practicó ensayo de tensión. Las probetas fueron caracterizadas mediante microscopia óp...
Análisis económico práctico de los bienes y servicios generadores de ingresos gravados con el Imp... more Análisis económico práctico de los bienes y servicios generadores de ingresos gravados con el Impuesto a la Renta, desde la perspectiva del uso, goce y utilidad de los mismos.
Canadian Mathematical Bulletin, 2015
A surface ∑ endowed with a Poisson tensor π is known to admit a canonical integration, 𝒢(π), whic... more A surface ∑ endowed with a Poisson tensor π is known to admit a canonical integration, 𝒢(π), which is a 4-dimensional manifold with a (symplectic) Lie groupoid structure. In this short note we show that if π is not an area form on the 2-sphere, then 𝒢(π) is diffeomorphic to the cotangent bundle T*∑. This extends results by the author and by Bonechi, Ciccoli, Staffolani, and Tarlini.
Indagationes Mathematicae, 2014
We construct a corank one Poisson manifold which is of strong compact type, i.e., the associated ... more We construct a corank one Poisson manifold which is of strong compact type, i.e., the associated Lie algebroid structure on its cotangent bundle is integrable, annd the source 1-conected (symplectic) integration is compact. The construction relies on the moduli of marked K3 surfaces.
Proceedings of the 30th annual international ACM SIGIR conference on Research and development in information retrieval, 2007
Query-by-semantic-description (QBSD) is a natural paradigm for retrieving content from large data... more Query-by-semantic-description (QBSD) is a natural paradigm for retrieving content from large databases of music. A major impediment to the development of good QBSD systems for music information retrieval has been the lack of a cleanlylabeled, publicly-available, heterogeneous data set of songs and associated annotations. We have collected the Computer Audition Lab 500-song (CAL500) data set by having humans listen to and annotate songs using a survey designed to capture 'semantic associations' between music and words. We adapt the supervised multi-class labeling (SML) model, which has shown good performance on the task of image retrieval, and use the CAL500 data to learn a model for music retrieval. The model parameters are estimated using the weighted mixture hierarchies expectation-maximization algorithm which has been specifically designed to handle realvalued semantic association between words and songs, rather than binary class labels. The output of the SML model, a vector of class-conditional probabilities, can be interpreted as a semantic multinomial distribution over a vocabulary. By also representing a semantic query as a query multinomial distribution, we can quickly rank order the songs in a database based on the Kullback-Leibler divergence between the query multinomial and each song's semantic multinomial. Qualitative and quantitative results demonstrate that our SML model can both annotate a novel song with meaningful words and retrieve relevant songs given a multi-word, text-based query.
Proceedings of the 24th international conference on Machine learning, 2007
Eigenvalue problems are rampant in machine learning and statistics and appear in the context of c... more Eigenvalue problems are rampant in machine learning and statistics and appear in the context of classification, dimensionality reduction, etc. In this paper, we consider a cardinality constrained variational formulation of generalized eigenvalue problem with sparse principal component analysis (PCA) as a special case. Using 1-norm approximation to the cardinality constraint, previous methods have proposed both convex and non-convex solutions to the sparse PCA problem. In contrast, we propose a tighter approximation that is related to the negative log-likelihood of a Student's t-distribution. The problem is then framed as a d.c. (difference of convex functions) program and is solved as a sequence of locally convex programs. We show that the proposed method not only explains more variance with sparse loadings on the principal directions but also has better scalability compared to other methods. We demonstrate these results on a collection of datasets of varying dimensionality, two of which are high-dimensional gene datasets where the goal is to find few relevant genes that explain as much variance as possible.
Pacific Journal of Mathematics, 2013
A class of codimension-one foliations has been recently introduced by imposing a natural compatib... more A class of codimension-one foliations has been recently introduced by imposing a natural compatibility condition with a closed maximally nondegenerate 2-form. In this paper we study for such foliations the information captured by a Donaldson-type submanifold. In particular we deduce that their leaf spaces are homeomorphic to leaf spaces of 3-dimensional taut foliations. We also introduce surgery constructions to show that this class of foliations is broad enough. Our techniques come mainly from symplectic geometry.
Machine Learning, 2010
Generalized eigenvalue (GEV) problems have applications in many areas of science and engineering.... more Generalized eigenvalue (GEV) problems have applications in many areas of science and engineering. For example, principal component analysis (PCA), canonical correlation analysis (CCA) and Fisher discriminant analysis (FDA) are specific instances of GEV problems, that are widely used in statistical data analysis. The main contribution of this work is to formulate a general, efficient algorithm to obtain sparse solutions to a GEV problem. Specific instances of sparse GEV problems can then be solved by specific instances of this algorithm. We achieve this by solving the GEV problem while constraining the cardinality of the solution. Instead of relaxing the cardinality constraint using a 1-norm approximation, we consider a tighter approximation that is related to the negative log-likelihood of a Student's t-distribution. The problem is then framed as a d.c. (difference of convex functions) program and is solved as a sequence of convex programs by invoking the majorizationminimization method. The resulting algorithm is proved to exhibit global convergence behavior, i.e., for any random initialization, the sequence (subsequence) of iterates generated by the algorithm converges to a stationary point of the d.c. program. Finally, we illustrate the merits of this general sparse GEV algorithm with three specific examples of sparse GEV problems: sparse PCA, sparse CCA and sparse FDA. Empirical evidence for these examples suggests that the proposed sparse GEV algorithm, which offers a general framework Editors: Süreyya Özöǧür-Akyüz, Devrim Ünay, and Alex Smola.
Journal of the London Mathematical Society, 2004
For any closed oriented manifold M , the top degree multi-vector fields transverse to the zero se... more For any closed oriented manifold M , the top degree multi-vector fields transverse to the zero section of ∧ top T M are classified, up to orientation preserving diffeomorphism, in terms of the topology of the arrangement of its zero locus and a finite number of numerical invariants. The group governing the infinitesimal deformations of such multi-vector fields is computed, and an explicit set of generators exhibited. For the sphere S n , a correspondence between certain isotopy classes of multi-vector fields and classes of weighted signed trees is established.
Infection and Immunity, 2004
ABSTRACTThe control ofListeria monocytogenesinfection depends on the rapid activation of the inna... more ABSTRACTThe control ofListeria monocytogenesinfection depends on the rapid activation of the innate immune system, likely through Toll-like receptors (TLR), since mice deficient for the common adapter protein of TLR signaling, myeloid differentiation factor 88 (MyD88), succumb toListeriainfection. In order to test whether TLR2 is involved in the control of infections, we compared the host response in TLR2-deficient mice with that in wild-type mice. Here we show that TLR2-deficient mice are more susceptible to systemic infection byListeriathan are wild-type mice, with a reduced survival rate, increased bacterial burden in the liver, and abundant and larger hepatic microabscesses containing increased numbers of neutrophils. The production of tumor necrosis factor, interleukin-12, and nitric oxide and the expression of the costimulatory molecules CD40 and CD86, which are necessary for the control of infection, were reduced in TLR2-deficient macrophages and dendritic cells stimulated by...
Immunity, 2007
Invariant natural killer T (iNKT) cells are a subset of innate lymphocytes that recognize lipid a... more Invariant natural killer T (iNKT) cells are a subset of innate lymphocytes that recognize lipid antigens in the context of CD1d and mediate potent immune regulatory functions via the rapid production of interferon-g (IFN-g) and interleukin-4 (IL-4). We investigated whether diverse Toll-like receptor (TLR) signals in myeloid dendritic cells (DCs) could differentially stimulate iNKT cells. Together with the lipopolysaccharide-detecting receptor TLR4, activation of the nucleic acid sensors TLR7 and TLR9 in DCs were particularly potent in stimulating iNKT cells to produce IFN-g, but not IL-4. iNKT cell activation in response to TLR9 stimulation required combined synthesis of type I interferon and de novo production of charged blinked glycosphingolipid(s) by DCs. In addition, DCs stimulated via TLR9 activated both iNKT cells and NK cells in vivo and protected mice against B16F10-induced melanoma metastases. These data underline the role of TLR9 in iNKT cell activation and might have relevance to infectious diseases and cancer.
IEEE Transactions on Audio, Speech, and Language Processing, 2008
We present a computer audition system that can both annotate novel audio tracks with semantically... more We present a computer audition system that can both annotate novel audio tracks with semantically meaningful words and retrieve relevant tracks from a database of unlabeled audio content given a text-based query. We consider the related tasks of content-based audio annotation and retrieval as one supervised multiclass, multilabel problem in which we model the joint probability of acoustic features and words. We collect a data set of 1700 human-generated annotations that describe 500 Western popular music tracks. For each word in a vocabulary, we use this data to train a Gaussian mixture model (GMM) over an audio feature space. We estimate the parameters of the model using the weighted mixture hierarchies expectation maximization algorithm. This algorithm is more scalable to large data sets and produces better density estimates than standard parameter estimation techniques. The quality of the music annotations produced by our system is comparable with the performance of humans on the same task. Our "query-by-text" system can retrieve appropriate songs for a large number of musically relevant words. We also show that our audition system is general by learning a model that can annotate and retrieve sound effects.
Comptes Rendus Mathematique, 2004
We prove that for closed 2-calibrated manifolds there always exist Lefschetz pencil structures. T... more We prove that for closed 2-calibrated manifolds there always exist Lefschetz pencil structures. This generalizes similar results for symplectic and contact manifolds. To cite this article: A.
Comptes Rendus Mathematique, 2004
The notion of 2-calibrated structure, generalizing contact structures, smooth taut foliations, et... more The notion of 2-calibrated structure, generalizing contact structures, smooth taut foliations, etc., is defined. Approximately holomorphic geometry as introduced by S. Donaldson for symplectic manifolds is extended to 2-calibrated manifolds. An estimated transversality result that enables to study the geometry of such manifolds is presented. To cite this article: A. Ibort, D.
Applied Physics Letters, 2014
Annals of Global Analysis and Geometry, 2011
We obtain universal models for several types of locally conformal symplectic manifolds via pullba... more We obtain universal models for several types of locally conformal symplectic manifolds via pullback or reduction. The relation with recent embedding results for locally conformal Kähler manifolds is discussed.
Journal of Symplectic Geometry, 2004
Uploads
Papers by DAVID MONTALUISA TORRES