Luíño Seoane – Department of Physics, Massachusetts Institute of Technology, USALanguage can be described using networks of (semanticaly, syntacticaly, …) interacting objects. Besides, a general information theory approach incorporates a speaker, a hearer, and a noisy channel. A key, common element in such approach is a confusion matrix encoding which words name each one of the existing objects, naturally introducing networks again. This also allows us to measure costs associated to communication across the channel for hearers and speakers. A rich literature on least-effort language exploring the optimality of communication codes developed from these methods. However, no systematic analysis of the underlying landscape of language graphs has been performed. We do this here, finding a rather complex and heterogeneous morphospace of language networks. We also derive a series of results relevant for the least-effort study of human language, largely in contradiction with theoretical speculations in the literature. These analysis are complemented, for the first time, with an empirical study of English words. Based on the WordNet database, we locate English vocabulary within the language networks morphospace. The outcome of this empirical analysis stresses the role of referential particles for efficient communication and to explore this morphospace.
Luíño Seoane studied Physics at the Universidade de Santiago de Compostela and Computational Neuroscience at the Technische Universität Berlin before completing his PhD in Complex Systems at the Pompeu Fabra University. There, under the supervision of Ricard Solé, he studied Multiobjective Optimization, its connections to Statistical Mechanics, and how optimization trade-offs lead to phase transitions and other phenomenology in complex networks, in linguistics, and in Darwinian evolution. His research in Neuroscience spans from Brain-Computer Interfaces to correlates of consciousness, while he is interested in developing research in Machine Learning and models of spiking neurons.
In his presentation, he will describe the current state of AIT’s telehealth projects for diabetes (DiabMemory) and cardiac (HerzMobil) telemonitoring for chronically ill patients. AIT’s telehealth projects are already running in routine care in Austria, financed within the Austrian healthcare system, and they have been pilots for Austria’s currently developing tele-health-service standards, which are intended to link telehealth to the Austrian Electronic Health Record ELGA.
Dieter Hayn will also present current results of their Predictive Analytics Toolset for Healthcare (PATH), which supports the rapid setup of predictive modelling solutions for various applications. Results from different applications of PATH will be discussed, including prediction of Delirium at hospital admission; blood transfusion needs prior surgery, re-admission at discharge, etc. A current application of PATH in a real-world clinical scenario at a hospital in Graz (Austria) will be presented, including visualization tools for plausibility checks of machine learning results.
Seminario impartido por los profesores Anna Rising y Jan Johansson, del Instituto Karolinska, (Estocolmo, Suecia)
Ponencia organizada por Brainvestigations, empresa española que estudia las ondas cerebrales con un objetivo claro: aplicar y llevar el conocimiento científico al mundo de los negocios.
Ciencia, employer branding, aplicaciones neurocientíficas para validar políticas de recursos humanos, y cómo las técnicas de neuroimagen pueden decodificar la mente para aportar datos muy importantes que en estos momentos es más difícil obtener de otro modo, son otros de los temas que se pretenden abordar en esta jornada
En el encuentro tomarán parte Fernando Maestú, catedrático y director del Laboratorio Neurociencia Cognitiva del CTB e Ignacio Belinchón, consultor de Recursos Humanos y consejero de Brainvestigations.
Para asistir al evento es necesario completar el siguiente formulario: www.brainvestigations.com/contacto
Más información, firstname.lastname@example.org
Tenemos el placer de invitarles a la conferencia que tendrá lugar el día 02 de marzo a las 12:00 en el Aula 01 de la planta cero del Centro de Tecnología Biomédica, impartida por el investigador Christophe Letellier.
The observability of a complex system refers to the property of being able to infer its whole state by measuring the dynamics of a limited set of its variables. Since in practice, monitoring all the variables defining the system’s state is experimentally unfeasible or inefficient, it is of utmost importance to develop a methodological framework addressing the problem of targeting those variables yielding full observability. Despite several approaches have been proposed, most of them neglect the nonlinear nature typically exhibited by complex systems and/or do not provide the space reconstructed from the measured variables. On the one hand, since nonlinearities are often related to a lack of observability, linear approaches cannot properly address this problem. On the other hand, finding the appropriate combination of sensors (and time derivatives) spanning the reconstructed space is a very time demanding computational task for large dimensional systems. Here, we adopt a nonlinear symbolic approach taking into account the nature of the interactions among variables and analyze the distribution of the linear and nonlinear load of the variables in the symbolic Jacobian matrix of the system. By means of two easy-to-implement criteria we are able to successfully identify the minimal set of variables (and their time derivatives) candidate to be measured for completing the reconstructed space. Our predictions are in full agreement with the analytical solution and drastically reduce the search for candidate variables, thus providing a key step to observe and model natural and man made complex systems of large dimension . Some explicit examples will be provided . If I have time, I will end with an application of observability for a follow-up in oncology.
Tenemos el placer de invitarles a la conferencia que tendrá lugar el día 27 de octubre a las 12:30 en el Salón de Actos del Centro de Tecnología Biomédica, impartida por el investigador de la Universidad de Wageningen, Sidharam P. Pujari.
CV S.P.Pujari: he received a B.Sc. in Chemistry at the Shivaji University, India and an M.Sc. at the University of Pune, India. He worked at the National Chemical Laboratory in Pune, India, and at the National Taiwan University Science and Technology, Taipei before starting a PhD in Prof. Zuilhofs labs in Wageningen University on covalently bound fluorinated monolayers. Subsequently he was a postdoc at the Zuilhof labs, and visiting scholar at the University of Texas at Dallas with Prof. Y. J. Chabal. Currently, he works as research associate position in Wageningen in the fields of materials and surface chemistry. He has published 40 articles with 2 (eu & us) patents with h index = 15.
Health prediction and treatment processes have become increasingly complex: involving huge amounts of data; with information coming from diverse sources; and the need of accessing to real-time data for decision-making. On the other hand, biocomputing has become a cornerstone of life sciences in the recent years, with high performance computing and analytics solutions as drivers of this digital transformation.
In this workshop, hosted by Atos Spain S.A. as leader of PAPHOS consortium, we will present PAPHOS platform offering, a holistic solution to address healthcare givers, clinical researchers and life science professionals’ problems for an effective, smooth and affordable translation of big data analytics to their practice.
Tenemos el placer de invitarles a la conferencia que tendrá lugar el día 16 de octubre a las 10:00 en el Salón de Actos del Centro de Tecnología Biomédica, impartida por el profesor James Yang.
Abstract: Digital human modeling and simulation plays an important role in product design,prototyping, manufacturing, sports biomechanics, and other areas. It reduces the number of design iterations and increases the safety and design quality of products. In this talk, I will first briefly review the state of art of digital human models, then I will summarize research projects carried out in my research lab. The first area is in engineering by investigating the optimization based digital human models to assist design and engineering, slips and falls. Applications include driver accommodation study and special population posture and motion prediction. The second area is healthcare engineering and spine biomechanics.