BDVA 2016 Valencia Summit

La Universitat Politècnica de València acogerá la tercera edición del Summit, organizado por la Big Data Value Association (BDVA), que tendrá lugar en Valencia del 29 de noviembre al 2 de diciembre.

Big Data Value Association es una asociación de la que forman parte los grandes de la industria de los macrodatos, como INDRA, IBM, ATOS, ORANGE o NOKIA, entre otros.

El objetivo de este Summit es dar a conocer las actividades llevadas a cabo desde la Asociación, presentar las oportunidades de financiación y servir de punto de encuentro para discutir retos y oportunidades en la European Big Data Community.
El evento está dirigido a empresas, centros de investigación y universidades, consorcios de organizaciones, tanto públicas como privadas, que desarrollen soluciones innovadoras en el área de Big Data o hayan incorporado con éxito, algunas de estas soluciones.

Durante el Summit se presentarán casos de éxito e innovadores y se darán las claves para que una organización de los primeros pasos para la adopción de estas tecnologías. Además, contará con un área de exhibición donde los principales actores en el área de Big Data tendrán la oportunidad de presentar sus soluciones a los asistentes, crear nuevas oportunidades de colaboración y organizar reuniones cara a cara.

Dentro del programa del evento, destacamos estas participaciones de miembros del PRHLT:

  • Conferencia “Deep Learning for Big Data” impartida por Roberto Paredes en sesión plenaria el día 1 de diciembre a las 18:00
  • Conferencia en la sesión paralela Deep Learning/Predictive Analysis impartida por Jon Ander Gómez y Roberto Paredes el día 1 de diciembre a las 12:15

Más información del Summit

Programa del Summit

Solver ML – Nueva spin off dedicada al desarrollo de modelos para el análisis de datos

La Universitat Politècnica de València (UPV) ha puesto en marcha recientemente una nueva spin off. Se trata de Solver Machine Learning, empresa de base tecnológica especializada en el desarrollo de modelos descriptivos y predictivos aplicados al análisis de datos, en particular al análisis de grandes volúmenes de datos (Big Data Analytics).

La empresa ha sido promovida por los investigadores del centro PRHLT, Jon Ander Gómez, Roberto Paredes y Francisco Casacuberta.

Más información

Vídeo de presentación

Gestures à Go Go – Aplicación que permite crear artificialmente escritura aparentemente humana

Daniel Martín-Albo (investigador del centro PRHLT) y Luis Leiva (antiguo investigador del PRHLT y socio fundador de Sciling) junto con investigadores de la École Polytechnique de Montréal (Canadá), han desarrollado Gestures à Go Go (G3), una nueva aplicación que permite crear artificialmente escritura aparentemente humana.
Su trabajo, publicado en la revista ACM Transactions on Intelligent Systems and Technology, ha sido recientemente demostrado durante la celebración del congreso MobileHCI.

Más información

Unsupervised Training Seminar

Next Wednesday, July 27 2016, at 12:30 in the Sala de Juntas of the DSIC, Professor Hermann Ney from RWTH Aachen, will give a talk on “Towards Unsupervised Training in Human Language Technology: Some Preliminary Results”

Abstract

The traditional approach to the training of statistical HLT systems (e.g. speech and handwriting recognition, machine translation) is to use annotated data, i.e. pairs of observation/output strings. E.g. in speech and handwriting recognition, for each string of observations, we need its orthographic transcription.

In contrast to this supervised training, this talk will focus on unsupervised training, for which no annotated training data is available. Instead we assume that a high-quality language model for the output strings is available. In addition, we are given a (very) long string of observations without any annotations.

For this type of unsupervised training, we will consider the resulting scientfic questions like: What is a suitable training criterion? What is the complexity of the training algoritm? How does the method compare with supervised training? We will discuss the state of the art and report on our ongoing work.

Seminario en HTR con la participación del Prof. Réjean Plamondon. lunes 11 de julio

El próximo 11 de julio a las 12:00, en la Sala INNOVA de la CPI, tendrá lugar un seminario sobre temas de procesado de texto manuscrito. En dicho seminario el Prof. Réjean Plamondon, de la École Polytechnique de Montréal impartirá una conferencia sobre las líneas de trabajo de su institución en estos temas.

Abstract:
Prof. Réjean Plamondon is a Full Professor in the department of Electrical Engineering at École Polytechnique de Montréal and Head of Laboratoire Scribens at this institution. Its institution is involved in many pattern recognition projects, particularly in the field of on-line and off-line handwriting analysis and processing. In this conference an overview of these projects will be given. Then, some points of common interest and general ideas for future collaboration with the PRHLT research center will be discussed.

Conferencia del Prof. Réjean Plamondon, viernes 8 de julio

El próximo 8 de julio a las 11:00, el Prof. Réjean Plamondon, de la École Polytechnique de Montréal impartirá en la Sala de Juntas del DSIC la conferencia titulada “The Lognormality principle”. El abstract se presenta a continuación.

Abstract:

The Kinematic Theory of rapid human movements and its family of lognormal models provide analytical representations of pen tip strokes, often considered as the basic unit of handwriting. This paradigm has not only been experimentally confirmed in numerous predictive and physiologically significant tests but it has also been shown to be the ideal mathematical description of the impulse response of a neuromuscular system. This proof has led to postulate the LOGNORMALITY PRINCIPLE. In its simplest form, this fundamental premise states that the lognormality of the neuromuscular impulse responses is the result of an asymptotic convergence, a basic global feature reflecting the behaviour of individuals who are in perfect control of their movements. As a corollary, motor control learning in young children can be interpreted as a migration toward lognormality. For the larger part of their lives, healthy human adults take advantage of lognormality to control their movements. Finally, as aging and health issues intensify, a progressive departure from lognormality is occurring. To illustrate this principle, we present various software tools and psychophysical tests used to investigate the fine motor control of subjects, with respect to these ideal lognormal behaviors, from childhood to old age. In this latter case, we focus particularly on investigations dealing with brain strokes, Parkinson and Alzheimer diseases. We also show how lognormality can be exploited in many pattern recognition applications for automatic generation of gestures, signatures, words and script independent patterns as well as CAPTCHA production, graffiti generation, anthropomorphic robot control and even speech modelling. Among other things, this lecture aims at elaborating a theoretical background for many handwriting applications as well as providing some basic knowledge that could be integrated or taking care of in the development of new automatic pattern recognition systems to be used for e-Learning, e-Security and e-Health.

PRHLT-Sciling collaboration awarded at Valencia IDEA competition

vlcidea15 award
vlcidea15 award

The awards ceremony of the Valencia IDEA competition has been celebrated today. The CORAL project, which was funded by the European Commission by means of the SME instrument, has been awarded with an honorable mention. CORAL is a collaboration between the PRHLT center and Sciling, an SME co-founded by current and past PRHLT members. The project aims to provide cheap and agile multilingualism for today’s websites.

Congratulations!

READ: Large European project on Handwritten Recognition evaluated with a score of 15/15

READ (Recognition and Enrichment of Archival Documents), a large European project proposal developed as a follow-up of the PRHLT-led tranScriptorium project, has recently been favorably evaluated by the Commission with the maximum evaluation score of 15 over 15.

READ was presented to the H2020-EINFRA-2015-1 call by a multidisciplinary consortium of 13 partners working in Computer Science, Pattern Recognition, Machine Learning, Image Processing and Humanities. It is coordinated by Günter Muehlberger from University of Innsbruck (Austria) and PRHLT is one of the leading partners concerning research in Handwritten Text Recognition.

READ will run for 3.5 years, with a total budget of exceeding 8 million Euros. The negotiation phase is now being finished and the consortium plans to start working in January, 2016.