# Liste de nos séminaires

*(ordre anti-chronologique)*

**Séminaire DAPA** du** 15 / 5 / 2014** à **10h**

*A normal hierarchical model for random intervals / The silhouette index - an extension to fuzzy clustering and applications to feature selection*

Dan Ralescu / Anca Ralescu

*Department of Mathematical Sciences, University of Cincinnati, USA / Computer Sciences, EE & CS Dept. College of Engineering University of Cincinnati, USA*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

(10h30-11h20) Dan Ralescu, Professor, Department of Mathematical Sciences, University of Cincinnati, USA*A normal hierarchical model for random intervals*

Existing methods for analyzing interval-valued data include regressions in the metric space of intervals and symbolic data analysis, the latter being proposed in a more general setting. However, there has been a lack of literature on the parametric modeling and distribution-based inferences for interval-valued data.

(11h20-12h10) Anca Ralescu, Professor, Computer Sciences, EE & CS Dept. College of Engineering University of Cincinnati, USA*The silhouette index - an extension to fuzzy clustering and applications to feature selection*

**Séminaire DAPA** du** 20 / 2 / 2014** à **10h**

*Robust recommendations and their explanation in multi-criteria decision aiding*

Christophe Labreuche

*Thales Group, France*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

Multi-Criteria Decision Aid (MCDA) aims at helping an individual to make choices among alternatives described by several attributes, from a (small) set of learning data representing her preferences. MCDA has a wide range of applications in smart cities, engineering, recommender systems and so on. Among the variety of available decision models, one can cite the weighted majority, additive utility, weighted sum or the Choquet integral.

Once the expression of the decision model has been chosen, the generation of choices among alternatives is classically done as follows. In a constraint approach, from a set of learning data (representing for instance comparisons of alternatives), one then looks for the value of the model parameters compatible with the learning data, which maximizes some functional, e.g. an entropy or a separation variable on the learning data. The comparisons among alternatives are then obtained by applying the model with the previously constructed parameters. The major difficulty the decision maker faces is that there usually does not exist one unique value of the parameters compatible with the learning data. Hence this approach introduces much arbitrariness since the generated preferences are much stronger than the learning data.

Robust preference relations have been recently introduced in MCDA to overcome this difficulty. An alternative is said to be necessarily preferred to another one if the first one dominates the second for any value of the parameters compatible with the learning data. In Artificial Intelligence, this operator is often called entailment. It is actually a closure operator. This necessity preference relation is usually incomplete, unless the model is completely specified from the preferential information of the decision maker.

The introduction of robust preference relation brings many new challenges:

- algorithmic aspects: how to design efficient algorithms to construct it?
- explanation: how to explain to the decision maker the recommended robust preferences? In other words, how are the recommendations derived from the learning data?

We will address these points in the talk.

**Séminaire DAPA** du** 6 / 2 / 2014** à **10h**

*Apprentissage actif en classification évidentielle sous contraintes*

Violaine Antoine

*ISIMA Limos*

Lieu : salle 101, couloir 25-26, 4 place Jussieu, 75005 Paris

La classification évidentielle et non supervisée se caractérise par l'utilisation de fonctions de croyance, et notamment l'utilisation de la notion de partition crédale. Cette notion élargit le concept de partition nette, floue, probabiliste ou possibiliste. Ainsi, elle permet de mesurer de manière précise l'incertitude quant à l'affectation d'un objet à une classe.

La classification sous contraintes, également appelée classification semi-supervisée, est une approche qui introduit une connaissance a priori sous forme de contraintes sur la partition recherchée. Nous nous intéressons ici à des contraintes au niveau des objets : une contrainte Must-Link spécifie que deux objets doivent être dans la même classe alors qu'une contrainte Cannot-link indique que deux objets se trouvent dans des classes différentes. L'ajout de contraintes permet une amélioration sensible des résultats de classification. Néanmoins, dans le cadre d'applications réelles, il est parfois difficile d'obtenir un jeu de contraintes intéressant. L'apprentissage actif consiste donc à obtenir ces informations à moindre coût.

Dans cette présentation, nous proposons deux nouveaux algorithmes de classification sous contraintes utilisant le cadre théorique des fonctions de croyance. Grâce à la partition crédale qu'ils retournent, nous pouvons identifier de manière précise les objets problématiques pour la classification. Un nouvel algorithme d'apprentissage actif est alors proposé afin de réduire l'erreur de classification.

**Séminaire DAPA** du** 19 / 12 / 2013** à **10h**

*The raise of graph databases/dataspaces and their relations with Linked Data and Ontologies*

André Santanchè

*Universidade Estadual de Campinas, Brazil*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

**Séminaire DAPA** du** 16 / 12 / 2013** à **16h**

*Extended Logic Programming and Intelligent System Development*

Asushi INOUE

*University of Cincinnati*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

A long-term effort toward a general application framework

for intelligent systems is introduced. Many intelligent systems

adopt a knowledge-based system architecture, and their development

thus differs from other application development. Expressing knowledge

as rules shifts one's perspective from data manipulation to relation

investigation. Our recent progress about two components are focused

- Extended Logic Programming (ELP), i.e. the keystone of this framework,

and a multi-view visualization scheme in order to effectively and

efficiently visualize the reasoning processes of ELP. Few representative

applications are showcased as time allows.

**Reference**:

K. Springer, M. Henry, A. Inoue, "A General Application Framework for Intelligent Systems,"

The 20th Midwest Artificial Intelligence and Cognitive Science Conference (MAICS2009),

Fort Wayne, IN, pp. 188-195, 2009.

**Séminaire DAPA** du** 5 / 12 / 2013** à **10h**

*Granular Models for Time Series Forecasting*

Rosangela Ballini

*Institute of Economics, University of Campinas, Brazil*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

time series forecasting. These models are constructed in two phases.

The first one uses the clustering algorithms to find group structures in a

historical database. Two different approaches are discussed: fuzzy c-means

clustering and participatory learning algorithms. Fuzzy c-mean clustering,

which is a supervised clustering algorithm, is used to explore similar

data characteristics, such as trend or cyclical components. Participatory

learning induces unsupervised dynamic fuzzy clustering algorithms and

provides an effective alternative to construct adaptive fuzzy systems.

In the second phase, two cases are considered. In the first case, a

regression model is adjusted for each cluster and forecasts are produced

by a weighted combination of the local regression models. In the second

case, prediction data are classified according to the group structure

found in the database. Then, forecasts are produced using the cluster

centers weighted by the degree with which prediction data match the

groups. The weighted combination of local models constitutes a forecasting

approach called granular functional forecasting modeling, and the approach

based on weighted combination cluster centers comprises granular

relational forecasting modeling. The effectiveness of the granular

forecasting approaches is verified using three different applications:

average streamflow forecasting, pricing option estimation and modeling of

regime changes in Brazilian nominal interest rates.

**Séminaire DAPA** du** 7 / 11 / 2013** à **10h**

*Automated Feature Weighting in Naive Bayes for High-dimensional Data Classification*

Shengrui Wang

*Université de Sherbrooke*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

This talk is about our recent work in the area of feature weighting for high dimensional data classification (and clustering). The first part of my talk relates to Naive Bayes (NB for short) classifier. Currently, in many real-world applications, high-dimensionality poses a major challenge to conventional NB classifiers, due to noisy or redundant features and local relevance of these features to classes. In this work, we propose an automated feature weighting solution to enable the NB method to deal effectively with high-dimensional data. First a locally weighted probability model will be presented for implementing a soft feature selection scheme. Then an optimization algorithm will be presented to find the weights in linear time complexity, based on the Logitnormal priori distribution and the Maximum a Posteriori principle. Experimental studies will show the effectiveness and suitability of the proposed model for high-dimensional data classification.

In the second part of this talk, I will briefly present our work on central clustering of categorical data with automated feature weighting. A novel kernel-density-based definition of cluster center is proposed using a Bayes-type probability estimator. Then, an algorithm called k-centers is proposed incorporating a new feature weighting scheme by which each attribute is automatically assigned with a weight measuring its individual contribution for the clusters.

**Séminaire DAPA** du** 4 / 11 / 2013** à **10h**

*Abductive reasoning made easy with Prolog and Constraint Handling Rules*

Henning Christiansen

*Roskilde University*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

Abductive reasoning, or "abduction", means to find a best explanation for some unexpected observation. In a logical setting, an explanation can be a set of facts which, when added to our current knowledge base, makes it possible to prove the truth of the observation and, at the same time, is not inconsistent with the knowledge base. Abduction in this sense is a useful metaphor for many sorts of reasoning aiming at answering "why" or "what" questions such as medical diagnosis, language understanding and decoding of biological sequence data. Furthermore, models of abductive reasoning can lead to practical implementation techniques.

Introduced by Peirce, the notion has attracted much attention in philosophy, detective stories and computer science, most notably in logic programming. Until the shift of the millennium, abduction in logic programming was realized through complex meta-interpreters written in Prolog, which may have led to a view of abduction as being some hairy, difficult stuff, far too inefficient for any realistic applications. In this talk, we demonstrate how a fairly powerful version of abductive reasoning can be exercised through a direct use of Prolog, using its extension by Constraint Handling Rules as the engine to take care of abducible hypotheses.

**Séminaire DAPA** du** 10 / 10 / 2013** à **15h**

*New Perspectives in Social Data Management / Understanding Similarity Metrics in Neighbour-based Recommender Systems*

Sihem Amer-Yahia / Arjen de Vries

*Laboratoire d'Informatique de Grenoble / Centrum Wiskunde & Informatica, Amsterdam*

Lieu : salle 105, couloir 25-26, 4 place Jussieu, 75005 Paris

*New Perspectives in Social Data Management*

The web has evolved from a technology platform to a social milieu where a mix of factual, opinion and behavior data interleave. A number of social applications are being built to analyze and extract value from this data and is encouraging us to do data-driven research.

I will describe a perspective on why and how social data management is fundamentally different from data management as it is taught in school today. More specifically, I'll talk about social data preparation, social data exploration and social application validation.

This talk is based on published and ongoing work with colleagues at LIG, UT Austin, U. of Trento, U. of Tacoma, and Google Research.

*Understanding Similarity Metrics in Neighbour-based Recommender Systems*

Recommender systems aim to predict the content that a user would like based on observations of the online behaviour of its users. Research in the CWI Information Access group addresses different aspects of this problem, varying from how to measure recommendation results, how recommender systems relate to information retrieval models, and how to build effective recommender systems (note: we won the ACM RecSys 2013 News Recommender Systems challenge!). We would like to develop a general methodology to diagnose weaknesses and strengths of recommender systems. In this talk, I discuss the initial results of an analysis of the core component of collaborative filtering recommenders: the similarity metric used to find the most similar users (neighbours) that will provide the basis for the recommendation to be made. The purpose is to shed light on the question why certain user similarity metrics have been found to perform better than others. We have studied statistics computed over the distance distribution in the neighbourhood as well as properties of the nearest neighbour graph. The features identified correlate strongly with measured prediction performance - however, we have not yet discovered how to deploy this knowledge to actually improve recommendations made.

**Séminaire DAPA** du** 19 / 9 / 2013** à **15h**

*Fuzzy Semantic Sentence Similarity Measures *

Keeley A Crockett

*The Intelligent Systems Group, School of Computing, Maths and Digital Technology, Manchester Metropolitan University*

Lieu : 25-26:105

A problem in the field of semantic sentence similarity is the inability of sentence similarity measures to accurately represent perception based (fuzzy) words that are commonly used in natural language. Given the wide use of fuzzy words in natural language this limits the strength of these measures in the areas where they are practically applied.

This talk briefly reviews traditional semantic word and sentence similarity measures and then describes a new fuzzy measure known as FAST (Fuzzy Algorithm for Similarity Testing). FAST is an ontology based similarity measure that uses concepts of fuzzy logic and computing with words to allow for the accurate representation of fuzzy based words. Through empirical human experimentation fuzzy sets were created for six categories of words based on their levels of association with particular concepts. These fuzzy sets were then defuzzified and the results used to create new ontological relations between the fuzzy words. These relationships allowed for the creation of a new ontology based semantic text similarity algorithm that is able to show the effect of fuzzy words on computing sentence similarity as well as the effect that fuzzy words have on non-fuzzy words within a sentence. Initial experiments using FAST are described on two possible future benchmark “fuzzy” datasets. The results show that there was an improved level of correlation between FAST and human test results compared with two traditional sentence similarity measures.

The talk concludes by looking at one potential application area where semantic similarity measures are utilised in a Student Debt Advisor Conversational Agent to remove the need for extensive scripting and maintenance.

**Séminaire DAPA** du** 19 / 7 / 2013** à **14h**

*Applications of a new effort based model of software usability*

**Dan E. Tamir**

*Texas State University, San Marcos, Texas*

Lieu : 26-00:101

Résumé :

The effort-based model for software usability stems from the notion that usability is an inverse function of effort. This new model of usability can be used for evaluating user interface, development of usable software, and pinpointing software usability defects. In this presentation, the underlying theory of the effort-based model along with pattern recognition techniques are used to introduce a framework for pinpointing usability deficiencies in software via automatic classification of segments of video file containing eye tracking results. In addition, we demonstrate the way that these principles can be used to construct a nondestructive user interface where the user can effectively navigate the web with minimum attention. The approach presented enables deriving web browsers for vehicle drivers and potentially for the blind.

Biographie :

Dr. Tamir is an associate professor in the Department of Computer Science, Texas State University, San Marcos, Texas (2005 - to date). He obtained the PhD-CS from Florida State University in1989, and the MS/BS-EE from Ben-Gurion University, Israel.

From 1996-2005, he managed applied research and design in DSP Core technology in Motorola SPS. From 1989-1996, he served as an assistant/associate professor in the CS Department at Florida Tech. Between 1983-1986, he worked in the applied research division, Tadiran, Israel.

Dr. Tamir is conducting research in combinatorial optimization, computer vision, audio, image, and video compression, human computer interaction, and pattern recognition.

**Séminaire DAPA** du** 28 / 6 / 2013** à **10h**

*Modeling Topics and Opinions in Asynchronous Conversations*

**Giuseppe Carenini**

*Department of Computer Science, University of British Columbia*

Lieu : Salle Champarnaud 26-00:124

ABSTRACT:

Due to the Internet revolution, human conversational data--in written forms--are accumulating at a phenomenal rate, as more and more people engage in email exchanges, blogging, texting and other social media activities. In this talk, we will present automatic methods for analyzing conversational text generated in asynchronous conversations, i.e., where participants communicate with each other at different times (e.g., email, blog, forum). Our focus will be on novel techniques to detect the topics covered in the conversation, and to identify whether an utterance in the conversation is expressing an opinion and what its polarity is.

Giuseppe Carenini, Associate Professor

Department of Computer Science

University of British Columbia

carenini@cs.ubc.ca, http://www.cs.ubc.ca/~carenini

BIO:

Giuseppe is an Associate Professor in Computer Science at the University of British Columbia (BC, Canada). He is also a member of the UBC Institute for Computing, Information, and Cognitive Systems (ICICS) and an Associate member of the UBC Institute for Resources, Environment and Sustainability (IRES). Giuseppe has broad interdisciplinary interests. His work on natural language processing and information visualization to support decision making has been published in over 80 peer-reviewed papers. Dr. Carenini was the area chair for “Sentiment Analysis, Opinion Mining, and Text Classification” of ACL 2009 and the area chair for “Summarization and Generation” of NAACL 2012. He has recently co-edited an ACM-TIST Special Issue on “Intelligent Visual Interfaces for Text Analysis”. In July 2011, he has published a co-authored book on “Methods for Mining and Summarizing Text Conversations”. In his work, Dr. Carenini has also extensively collaborated with industrial partners, including Microsoft and IBM. Giuseppe was awarded a Google Research Award and an IBM CASCON Best Exhibit Award in 2007 and 2010 respectively.

**Séminaire DAPA** du** 27 / 6 / 2013** à **10h**

*Exploration and Exploitation of Scratch Games*

Raphaël Féraud

*Orange Labs*

Lieu : 25-26:105

We consider a variant of the multi-armed bandit model, which we call scratch games, where the sequences of rewards are finite and drawn in advance with unknown starting dates. This new problem is motivated by online advertising applications where the number of ad displays is fixed according to a contract between the advertiser and the publisher, and where a new ad may appear at any time. The drawn-in-advance assumption is natural for the adversarial approach where an oblivious adversary is supposed to choose the reward sequences in advance. For the stochastic setting, it is functionally equivalent to an urn where draws are performed without replacement. The non-replacement assumption is suited to the sequential design of non-reproducible experiments, which is often the case in real world. By adapting the standard multi-armed bandit algorithms to take advantage of this setting, we propose three new algorithms: the first one is designed for adversarial rewards; the second one assumes a stochastic urn model; and the last one is based on a Bayesian approach. For the adversarial and stochastic approaches, we provide upper bounds of the regret which compare favorably with the ones of Exp3 and UCB1. We also confirm experimentally that these algorithms compare favorably with Exp3, UCB1 and Thompson Sampling by simulation with synthetic models and ad-serving data.

Keywords adversarial multi-armed bandits ; stochastic multi-armed bandits ; finite sequences ; scratch games.

**Séminaire DAPA** du** 13 / 6 / 2013** à **10h**

*Vers la gestion de l'imprécision dès la construction de systèmes d'information géographique à la visualisation des données : une démarche basée sur la théorie des ensembles flous*

Cyril de Runz

*CReSTIC, IUT de Reims Châlons Charleville*

Lieu : 25-26:105

Ce travail se positionne dans le cadre de la manipulation de données spatiotemporelles réelles en tenant compte de leur imperfection et plus particulièrement de leur imprécision. La démarche présentée s'inscrit dans la volonté d'aller vers une meilleure gestion de celles-ci tant pour leur représentation que pour leur analyse et leur visualisation. Dans ce contexte, nos contributions portent tant au niveau conception et construction des systèmes d'information géographique que de l'interrogation et l'exploration (possiblement visuelle) des données. Nos méthodes se basent notamment sur la définition de pictogrammes visuels étendant l'UML, d'indices temporels flous, de graphes, de rangs, de coloriage guidé par les données, etc. La démarche sera illustrée autour d'applications sur des données issues de l'archéologie préventive et sur des cas d'études prospectifs en agronomie et en urbanisme.

**Séminaire DAPA** du** 6 / 6 / 2013** à **10h**

*Experiments with Probabilistic Logic Programming applied to Biological Sequence Analysis*

Ole Torp Lassen

*Roskilde University, Danemark - LIP6 depuis 01/04/2013*

Lieu : 25-26:105

Systems that combine logic programming and statistical inference in theory allow machine learning systems to deal with both relational and statistical information.In practice, however, such applications do not scale very well.The LoSt project was concerned with a compositional approach to overcome those challenges. In particular, we experimented with applying one probabilistic logic programming system, PRISM (Taisuke Sato & Yoshitaka Kameya), based on B-Prolog, to complex, large scale bio-informatical problems.Firstly, some important aspects of the PRISM system and its underlying implementation were optimised for application to large scale data.Secondly, we developed a compositional method of analysis, Bayesian Annotation Networks, where the complex overall task is approximated by identifying and negotiating interdependent constituent subtasks and, in turn, integrating their analytical results according to their interdependencies.Finally, we experimented extensively with the developed framework in the domain of procaryotic gene-finding. As part of the general domain of DNA-annotation, the task of gene-finding is characterized by large sets of extremely long and highly ambiguous sequences of data and, thus, represents a suitably challenging setting for efficient analysis.In general, we concluded that with the computing power of today, probabilistic logic programming systems, as exemplified by PRISM, can be applied efficiently - also in large scale domains. As such, probabilistic logic programming offers extremely expressive models with very clear semantics – facilitating increased focus on domain properties and less on programming complexity.

**Séminaire DAPA** du** 30 / 5 / 2013** à **10h30**

*Exploring Categories of Uncertainty - toward Structure of Uncertainty*

Michio Sugeno

*Tokyo Institute of Technology, Japan and European Centre for Soft Computing, Spain*

Lieu : 25-26:105

As a conventional concept of uncertainty, we are familiar with the 'probability' of a phenomenon initiated in 17 century. Also we often discuss the 'uncertainty' of knowledge. Recently, Fuzzy Theory has brought a hidden uncertainty, 'fuzziness', to light. Reflections on these ideas lead to a fundamental question: What kinds of uncertainty are we aware of? Motivated by this question, this study aims to explore categories and modalities of uncertainty. For instance, we have found that (i) 'form' is a category of uncertainty; (ii) 'inconsistency' is a modality of uncertainty; (iii) the inconsistency of form is one of the major uncertainties. Through the classification of adjectives implying various uncertainties, we elucidate seven uncertainties (or nine if subcategories are counted) and identify three essential ones among them, such as the fuzziness of wording. Finally the structure of uncertainty will be shown. The obtained structure is verified by psychological experiments, while the validity of three essential uncertainties is examined by linguistic analysis.

**Séminaire DAPA** du** 16 / 5 / 2013** à **10h**

*Introduction to Active Sets*

Germano Resconi

*Department of Mathematics and Physics, Catholic University, Brescia, Italie*

Lieu : 25-26:105

An active set is a unifying space being able to act as a “bridge” for transferring information, ideas and results between distinct types of uncertainties and different types of applications. An active set is a set of agents who independently deliver true or false values for a given proposition. An active set is not a simple vector of logic values for different propositions, the results are a vector but the set is not.

The difference between an ordinary set and active set is that the ordinary set has passive elements with values of the attributes defined by an external agent, in the active set any element is an agent that internally defines the value of a given attribute for a passive element.

Agents in the active set with a special criteria gives the logic value for the same attribute. So agents in many cases are in a logic conflict and this generate semantic uncertainty on the logic evaluation. Criteria and agents are the two variables by which we give different logic values to the same attribute or proposition. Active sets is beyond the modal logic. In fact given a proposition in modal logic we can evaluate the proposition only when we know the worlds where the proposition is locate. When we evaluate one proposition in one world we cannot evaluate the same proposition in another world. Now in epistemic logic any world is an agent that know that the proposition is true or false. Now the active set is a set of agents as in the epistemic logic but the difference with modal logic is that all the agents (worlds) are not separate but are joined in the evaluation of the given proposition. In active set for one agent and one criteria we have one logic value but for many agents and criteria the evaluation is not true and false but is a matrix of true and false. This matrix is not only a logic evaluation as in the modal logic but give us the conflicting structure of the active set evaluation. Matrix agent is the vector subspace of the true false agent multi dimension space. Operations among active set include operations in the traditional set , fuzzy sets and rough set as special cases. The agent multi dimensional space to evaluate active set include also the Hilbert multidimensional space where is possible to simulate quantum logic gate. New logic operation are possible as fuzzy gate operations and more complex operations as conflicting solving , consensus operations , syntactic inconsistency , semantic inconsistency and knowledge integration. In the space of the agents evaluations morphotronic geometric operations are the new frontier to model new types of computers , new type of model for wireless communications as cognitive radio. In conclusion Active set open new possibility and new models for the logic.

**Séminaire DAPA** du** 15 / 5 / 2013** à **10h**

*Neuromuscular Modelling and Analysis of Handwriting: from Automatic Generation to Biomedical and Neurocognitive applications.*

Réjean Plamondon

*Laboratoire Scribens, Département de Génie Électrique, École Polytechnique de Montréal*

Lieu : 25-26:105

Many models have been proposed over the years to study human movements in general and

handwriting in particular: models relying on neural networks, dynamics models, psychophysical

models, kinematic models and models exploiting minimization principles. Among the models

that can be used to provide analytical representations of a pen stroke, the Kinematic Theory of

rapid human movements and its family of lognormal models has often served as a guide in the

design of pattern recognition systems relying on the exploitation of the fine neuromotricity, like

on-line handwriting segmentation, signature verification as well as in the design of intelligent

systems involving in a way or another, the global processing of human movements. 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 exploited in

biomedical engineering and cognitive neurosciences.

More specifically, we will overview the basic neuromotor properties of single strokes and will

explain how they can be superimposed vectorially to generate complex pen tip trajectories.

Doing so, we will report on various projects conducted by our team and our collaborators. First,

we will present a brief comparative survey of the different lognormal models. Then, from a

practical perspective, we will describe some parameter extraction algorithms suitable for the

reverse engineering of individual strokes as well as of complex handwriting signals. We will show

how the resulting representation could be employed to characterize signers and writers and

how the corresponding feature sets could be exploited to study the effects of various factors,

like aging and health problems, on handwriting variability. We will also describe some

methodologies to generate automatically huge on-line handwriting databases for either writer

dependent or writer independent applications as well as for the production of synthetic

signature databases. From a theoretical perspective, we will explain how, using an original

psychophysical set up, we have been able to validate the basic hypothesis of the Kinematic

Theory and to test its most distinctive predictions. We will complete this survey by explaining

how the Kinematic Theory could be utilized to improve some signal processing techniques,

opening a window on novel potential applications for on-line handwriting processing,

particularly to provide some benchmarks to analyze children handwriting learning, to study

aging effects on neuromotor control as well as developing diagnostic systems for

neuromuscular disorders. To illustrate this latter point, we will report typical results obtained so

far for the assessment of brain stroke most important modifiable risk factors (diabetes,

hypertension, hypercholesterolemia, obesity, cardiac problems, cigarette smoking).

**Séminaire DAPA** du** 11 / 4 / 2013** à **10h**

*Co-clustering sous différentes approches: Modèles et algorithmes*

Mohamed Nadif

*LIPADE, Université Paris Descarte*

Lieu : 25-26:105

La classification automatique est devenue un outil important qui s'est beaucoup développé ces dernières années. Bien que les procédures de classification soient nombreuses et que la majorité d'entre elles ait pour objectif de construire une partition optimale des objets (lignes) ou des variables (colonnes), il existe d'autres méthodes, dites de classification croisée, qui considèrent les deux ensembles simultanément et cherchent à organiser les données en blocs homogènes. Comparées aux méthodes de classification classiques, les algorithmes de classification croisée ont démontré leur efficacité dans la découverte de structures à partir de matrices de données de grande taille (lignes et/ou colonnes), sparses ou non. Dans ma présentation, je vais considérer plusieurs approches en insistant particulièrement sur l'approche mélange basée sur les modèles latents par blocs (Latent Block Models) et l'approche factorisation basée sur la tri-factorisation de matrice non négative (Nonnegative Matrix Tri-Factorization).

**Séminaire DAPA** du** 28 / 3 / 2013** à **11h**

*Spike-based computing and learning in brains, machines, and visual systems in particular*

Timothée Masquelier

*Adaptive NeuroComputation Group, Lab. of Neurobiology of Adaptive Processes, UPMC*

Lieu : 25-26:105

Using simulations, we have first shown that, thanks to the physiological learning mechanism referred to as Spike Timing-Dependent Plasticity (STDP), neurons can detect and learn repeating spike patterns, in an unsupervised manner, even when those patterns are embedded in noise[1,2]. Importantly, the spike patterns do not need to repeat exactly: it also works when only a firing probability pattern repeats, providing this profile has narrow (10-20ms) temporal peaks[3]. Brain oscillations may help in getting the required temporal precision[4], in particular when dealing with slowly changing stimuli. All together, these studies show that some envisaged problems associated to spike timing codes, in particular noise-resistance, the need for a reference time, or the decoding issue, might not be as severe as once thought. These generic STDP-based mechanisms are probably at work in particular the visualsystem, where they can explain how selectivity to visual primitives emerges[5,6], leading to very reactive systems. I am now investigating if they are also at work in the somatosensory system. Finally, these mechanisms are also appealing for neuromorphic engineering: they can be efficiently implemented in hardware, leading to fast systems with self-learning abilities[7].

References

1 Masquelier, T. et al. (2008) Spike timing dependent plasticity finds the start of repeating patterns in continuous spike trains. PLoS ONE 3, e1377

2 Masquelier, T. et al. (2009) Competitive STDP-Based Spike Pattern Learning. Neural Comput 21, 1259–1276

3 Gilson, M. et al. (2011) STDP allows fast rate-modulated coding with Poisson-like spike trains. PLoS Comput Biol 7, e1002231

4 Masquelier, T. et al. (2009) Oscillations, phase-of-firing coding, and spike timing-dependent plasticity: an efficient learning scheme. The Journal of neuroscience 29, 13484–93

5 Masquelier, T. and Thorpe, S.J. (2007) Unsupervised learning of visual features through spike timing dependent plasticity. PLoS Comput Biol 3, e31

6 Masquelier, T. (2012) Relative spike time coding and STDP-based orientation selectivity in the early visual system in natural continuous and saccadic vision: a computational model. Journal of computational neuroscience 32, 425–41

7 Zamarreño-Ramos, C. et al. (2011) On Spike-Timing-Dependent-Plasticity, Memristive Devices, and Building a Self-Learning Visual Cortex. Frontiers in neuroscience 5, 22