Differences
This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision | ||
public:seminars_manifestations [2015/04/03 10:57] – external edit 127.0.0.1 | public:seminars_manifestations [2017/03/08 09:17] (current) – fmichel | ||
---|---|---|---|
Line 1: | Line 1: | ||
- | **To announce | + | **[[http:// |
+ | |||
+ | |||
+ | |||
+ | **SPARKS public calendar**: in [[https:// | ||
+ | |||
+ | |||
+ | ====== 2017 ====== | ||
+ | |||
+ | ===== Séminaires ===== | ||
+ | |||
+ | ==== Marie-Christine ROUSSET: Datalog revisited for reasoning in Linked Data ==== | ||
+ | When: 3rd March 2017, 10h00 \\ | ||
+ | Where: POLYTECH, Templiers 2, room 307\\ | ||
+ | |||
+ | **Abstract**\\ | ||
+ | Linked Data provides access to huge, continuously growing amounts of open data and ontologies in RDF format that describe entities, links and properties on those entities. Equipping Linked Data with inference paves the way to make the Semantic Web a reality. In this presentation, | ||
+ | equipping RDF triple stores with Datalog inference rules. This rule language allows to capture in a uniform manner OWL constraints that are | ||
+ | useful in practice, such as property transtivity or symmetry, but also domain-specific rules with practical relevance for users in many domains of interest. I will illustrate the expressivity of this framework for modeling Linked Data applications and its genericity for developing inference algorithms. In particular, we will show how it allows to model the problem of data linkage in Linked Data as a reasoning problem on possibly decentralized data. I will also explain how it makes possible to efficiently extract expressive modules from Semantic Web ontologies and databases with formal guarantees, whilst effectively controlling their succinctness. Experiments conducted on real-world datasets have demonstrated the feasibility of this approach and its usefulness in practice for data integration and information extraction. | ||
+ | |||
+ | ===== Soutenances ===== | ||
+ | |||
+ | ==== PhD Thesis Defense - Franck MICHEL ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | === Abstract === | ||
+ | |||
+ | **Title: Integrating Heterogeneous Data Sources in the Web of Data** | ||
+ | |||
+ | To a great extent, RDF-based data integration as well as the Web of Data depend on the ability to reach out legacy data locked in data silos where they are invisible to the web. In the last 15 years, various works have tackled the problem of exposing structured data in the Resource Description Framework (RDF), starting with relational databases (RDB), spreadsheets and the XML data format. Meanwhile, the overwhelming success of NoSQL databases has made the database landscape more diverse than ever. So far, though, these databases remain inaccessible to RDF-based data integration systems, and although the data they host may be of interest to a large audience, they remain invisible to the Web of Data. Hence, to harness the potential of NoSQL databases and more generally non-RDF data sources, the objective of this thesis is to enable RDF-based data integration over heterogeneous databases and, in particular, to reconcile the Semantic Web with the NoSQL family of databases. | ||
+ | |||
+ | Firstly, we propose a generic mapping language, xR2RML, able to describe the mapping of varying types of databases into an arbitrary RDF representation. This language relies on and extends previous works on the translation of RDBs, CSV and XML into RDF. Secondly, we propose to use such an xR2RML mapping either to materialize RDF data or to dynamically evaluate SPARQL queries on the native database. To spur the development of SPARQL interfaces over legacy databases, we propose a two-step approach. The first step performs the translation of a SPARQL query into a pivot abstract query based on the xR2RML mapping of the target database to RDF. In the second step, the abstract query is translated into a concrete query, taking into account the specificities of the database query language. Great care is taken of the query optimization opportunities, | ||
+ | |||
+ | ====== 2016 ====== | ||
+ | |||
+ | ===== Séminaires ===== | ||
+ | |||
+ | ==== Matthieu Cord: Deep learning and weak supervision for image classification ==== | ||
+ | When: 17th May 2016, 10h00 \\ | ||
+ | Where: POLYTECH, room E+132\\ | ||
+ | By: Matthieu Cord | ||
+ | |||
+ | **Abstract**\\ | ||
+ | Deep learning and Convolutional Neural Networks (CNN) are state-of-the-art methods for various visual recognition tasks, e.g. image classification or object detection. To better identify or localize objects, bounding box annotations are often used. These rich annotations quickly become too costly to get, making the development of Weakly Supervised Learning (WSL) models appealing. | ||
+ | We discuss several strategies to automatically select relevant image regions from weak annotations (e.g. image-level labels) in deep CNN. We also introduce our architecture WELDON for WEakly supervised Learning of Deep cOnvolutional neural Networks. | ||
+ | Our deep learning framework, leveraging recent improvements on the Multiple Instance Learning paradigm, is validated on seceral recognition tasks. | ||
+ | |||
+ | ** About the speaker**\\ | ||
+ | Matthieu Cord is Professor in the Computer Science department LIP6, at UPMC University. | ||
+ | |||
+ | ** Slides **\\ | ||
+ | {{: | ||
+ | ==== Jörg Kienzle: Concern-Oriented Reuse ==== | ||
+ | When: 14th April 2016, 14h30 \\ | ||
+ | Where: EURECOM, room 101\\ | ||
+ | By: Jörg Kienzle | ||
+ | |||
+ | **Abstract** | ||
+ | |||
+ | MDE: In the context of MDE, my current research focusses on Concern-Oriented Reuse (CORE), a new software reuse paradigm in which software artifacts (models and/or code) are modularized by domains of abstraction within units of reuse called concerns. CORE combines techniques from software product line engineering and aspect-orientation to create extremely versatile, generic units of reuse with well-defined interfaces that encapsulate a variety of solutions to recurring software development problems and allow the user to reason about the impact of selecting a solution on high-level goals and system properties. I am going to give a Labex seminar on CORE this Thursday, April 14th, from 14:30 - 15:30 at EURECOM, room 101 (http://www.sophia-networking.org/ | ||
+ | While this talk is going to give a high-level overview on CORE, I am planning to present more details about CORE in the following weeks to the modelling groups at I3S, in particular on: | ||
+ | * Unordered List Item- how CORE can be used to simplify the use of existing code frameworks and APIs (paper at MODELS 2015) | ||
+ | * the algorithms and model transformations used in CORE and the TouchCORE tool | ||
+ | * how CORE addresses feature interactions, | ||
+ | |||
+ | I would be very interested in talking to anyone who is developing reusable libraries | ||
+ | |||
+ | Distributed Systems: In the context of distributed systems, I have been mainly working on data dissemination and fault tolerance in large-scale, | ||
+ | * " | ||
+ | * - " | ||
+ | * " | ||
+ | For all 3 papers, we ran extensive real-world experiments involving hundreds of machines and thousands of players on top of Mammoth, a massively multiplayer game research framework that we developed over the last 10 years. | ||
+ | |||
+ | ** About the speaker**\\ | ||
+ | Jörg Kienzle is an associate professor at the School of Computer Science at McGill University in Montreal, Canada, and is visiting the SPARKS team at I3S until end of June 2016. He is interested in meeting anyone at I3S / Inria / Eurecom with common interests to talk about potential collaboration. His research group works in two fields, model-driven engineering (MDE) and distributed systems. Description of the research topics and of a first seminar for the Labex on Thursday April 14th follow. | ||
+ | |||
+ | ===== Soutenances ===== | ||
+ | |||
+ | ==== PhD Thesis Defense - Atheer AL-NAJDI ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | === Abstract === | ||
+ | |||
+ | **Title: | ||
+ | |||
+ | Clustering is the process of partitioning a dataset into groups, so that the instances in the same group are more similar to each other than to instances in any other group. Many clustering algorithms were proposed, but none of them proved to provide good quality partition in all situations. Consensus clustering aims to enhance the clustering process by combining different partitions obtained from different algorithms to yield a better quality consensus solution. In this work, a new consensus clustering method, called MultiCons, is proposed. It uses the frequent closed itemset mining technique in order to discover the similarities between the different base clustering solutions. The identified similarities are presented in a form of clustering patterns, that each defines the agreement between a set of base clusters in grouping a set of instances. By dividing these patterns into groups based on the number of base clusters that define the pattern, MultiCons generates a consensus | ||
+ | solution from each group, resulting in having multiple consensus candidates. These different solutions are presented in a tree-like structure, called ConsTree, that facilitates understanding the process of building the multiple consensuses, | ||
+ | Five consensus functions are proposed in this work in order to build a consensus solution from the clustering patterns. Approach 1 is to just merge any intersecting clustering patterns. Approach 2 can either merge or split intersecting patterns based on a proposed measure, called intersection ratio. Approach 3 differs from the previous approaches by searching for the best similar pattern before making a merge/split decision, and, in addition, it uses the average intersection ratio. While approach 3 works sequentially on the clustering patterns, approach 4 uses a similarity matrix of intersection ratios to search for the best merge/ | ||
+ | |||
+ | **Keywords**: | ||
+ | |||
+ | === Résumé === | ||
+ | |||
+ | **Titre : Une approche basée sur les motifs fermés pour résoudre le problème de clustering par consensus** | ||
+ | |||
+ | Le clustering est le processus de partitionnement d’un ensemble de données en groupes, de sorte que les instances du même groupe sont plus semblables les unes aux autres qu’avec celles de tout autre groupe. De nombreux algorithmes de clustering ont été proposés, mais aucun d’entre eux ne s’avère fournir une partition des données pertinente dans toutes les situations. Le clustering par consensus vise à améliorer le processus de regroupement en combinant différentes partitions obtenues à partir de divers algorithmes afin d’obtenir une solution de consensus de meilleure qualité. Dans ce travail, une nouvelle méthode de clustering par consensus, appelée MultiCons, est proposée. Cette méthode utilise la technique d’extraction des itemsets fréquents fermés dans le but de découvrir les similitudes entre les différentes solutions de clustering dits de base. Les similitudes identifiées sont représentées sous une forme de motifs de clustering, chacun définissant un accord entre un ensemble de clusters de bases sur le regroupement d’un ensemble d’instances. En traitant ces motifs par groupes, en fonction du nombre de clusters de base qui définissent le motif, la méthode MultiCons génère une solution de consensus pour chaque groupe, générant par conséquence plusieurs consensus candidats. Ces différentes solutions sont ensuite représentées dans une structure arborescente appelée arbre de consensus, ou ConsTree. Cette représentation graphique facilite la compréhension du processus de construction des multiples consensus, ainsi que les relations entre les instances et les structures d’instances dans l’espace de données. | ||
+ | Cinq approches de clustering par consensus, permettant de construire une solution de consensus à partir des motifs de clustering, sont proposées dans ce travail. La première approche fusionne simplement successivement tous les motifs de clustering qui se recoupent. La seconde approche va soit fusionner, soit diviser les motifs qui se recoupent selon le résultat d’une nouvelle mesure appelée ratio d’intersection. La troisième approche diffère des approches précédentes en recherchant, | ||
+ | |||
+ | **Mots clés** : Clustering ; Classification non-supervisée ; Clustering par consensus ; Ensembles clustering ; Itemsets fréquents fermés. | ||
+ | |||
+ | |||
+ | ==== PhD Thesis Defense - Romaric Pighetti ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | === Abstract === | ||
+ | |||
+ | **Title: | ||
+ | |||
+ | Given the ever growing amount of visual content available on the Internet, the need for systems able to search through this content has grown. | ||
+ | Content based image retrieval systems have been developed to address this need. | ||
+ | But with the growing size of the databases, new challenges arise. | ||
+ | In this thesis, the fine grained classification problem is studied in particular. | ||
+ | It is first shown that existing techniques, and in particular the support vector machines which are one of the best image classification technique, have some difficulties in solving this problem. | ||
+ | They often lack of exploration in their process. | ||
+ | Then, evolutionary algorithms are considered to solve the problem, for their balance between exploration and exploitation. | ||
+ | But their performances are not good enough either. | ||
+ | Finally, an hybrid system combining an evolutionary algorithm and a support vector machine is proposed. | ||
+ | This system uses the evolutionary algorithm to iteratively feed the support vector machine with training samples. | ||
+ | The experiments conducted on Caltech-256, | ||
+ | |||
+ | **Keywords**: | ||
+ | |||
+ | |||
+ | === Résumé === | ||
+ | |||
+ | **Titre : Une méthode hybride pour la classification d' | ||
+ | |||
+ | La quantité d' | ||
+ | Les systèmes de recherche d' | ||
+ | Mais les bases de données grandissant, | ||
+ | Dans cette thèse, la classification à grain fin est étudiée en particulier. | ||
+ | Elle consiste à séparer des images qui sont relativement semblables visuellement mais représentent différents concepts, et à regrouper des images qui sont différentes visuellement mais représentent le même concept. | ||
+ | Il est montré dans un premier temps que les techniques classiques de recherche d' | ||
+ | Même les techniques utilisant les machines à vecteur de support (SVM), qui sont très performants pour la classification, | ||
+ | Ces techniques n' | ||
+ | D' | ||
+ | Toutefois, leurs performances restent encore limitées. | ||
+ | Par conséquent, | ||
+ | L' | ||
+ | Ce système est évalué avec succès sur la base de données Caltech-256 contenant envieront 30 000 images réparties en 256 catégories. | ||
+ | |||
+ | **Mots clés** : Recherche d' | ||
+ | |||
+ | ==== PhD Thesis Defense - Zide Meng ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | === Abstract === | ||
+ | **Title: Temporal and semantic analysis of richly typed social networks from user-generated content sites on the Web** | ||
+ | |||
+ | We propose an approach to detect topics, overlapping communities of interest, expertise, trends and activities in user-generated content sites and in particular in question-answering forums such as StackOverFlow. We first describe QASM (Question & Answer Social Media), a system based on social network analysis to manage the two main resources in question-answering sites: users and contents. We also introduce the QASM vocabulary used to formalize both the level of interest and the expertise of users on topics. We then propose an efficient approach to detect communities of interest. It relies on another method to enrich questions with a more general tag when needed. We compared three detection methods on a dataset extracted from the popular Q&A site StackOverflow. Our method based on topic modeling and user membership assignment is shown to be much simpler and faster while preserving the quality of the detection. We then propose an additional method to automatically generate a label for a detected topic by analyzing the meaning and links of its bag of words. We conduct a user study to compare different algorithms to choose the label. Finally we extend our probabilistic graphical model to jointly model topics, expertise, activities and trends. We performed experiments with real-world data to confirm the effectiveness of our joint model, studying the users’ behaviors and topics dynamics. | ||
+ | |||
+ | |||
+ | === Résumé === | ||
+ | |||
+ | **Titre : Analyse temporelle et sémantique des réseaux sociaux typés à partir du contenu de sites | ||
+ | généré par des utilisateurs sur le Web** | ||
+ | |||
+ | Nous proposons une approche pour détecter les sujets, les communautés d' | ||
+ | |||
+ | |||
+ | ==== PhD Thesis Defense - Papa Fary Diallo ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | === Abstract === | ||
+ | **Title: Sociocultural and Temporal Aspects in Ontologies dedicated to Virtual Communities** | ||
+ | |||
+ | **Keywords: Semantic web, Social web, Ontologies, Virtual Communities, | ||
+ | |||
+ | This thesis is set in a research effort that aims to model sociocultural and temporal aspects to allow Senegalese communities to share and to co-construct their sociocultural knowledge. Indeed, with the globalization it is very common to meet African youth and particularly Senegalese youth knowing more about the geography of the West than their own countries. Thus, to refresh the memory of our fellow citizens and revive the many stories that accompany the creation and daily life of the different Senegalese territories, | ||
+ | |||
+ | Our proposals are based on social and semantic web technologies. indeed, social web proposes a framework where value is created by the aggregation of many individual user contributions. Thereby, social web makes easier corpus co-construction. The semantic web enables to find, to combine and to share resources, not only between humans but also between machines. The combination of these two technologies enables Senegalese communities to share and co-construct their cultural heritage in a collaborative and semantic environment . | ||
+ | |||
+ | Our contributions include to (i) propose ontologies to annotate sociocultural resources and (ii) provide a framework for communities to share and co-construct their knowledge. Ontologies are backbone of the semantic web and allow to characterize a domain by describing the basic concepts and the relations between them. Thus, we have defined two ontologies : 1) a sociocultural ontology based on cultural-historical activity theory and 2) a temporal ontology to annotate temporally sociocultural resources. We also proposed a virtual community called cultural knowledge-building community which is an adaptation of the knowledge-building community in the cultural field. | ||
+ | |||
+ | |||
+ | === Résumé === | ||
+ | |||
+ | **Titre : Aspects Socioculturels et Temporels dans les Ontologies pour les Communautés Virtuelles | ||
+ | résumé** | ||
+ | |||
+ | **Mots clés : Web sémantique, | ||
+ | |||
+ | Les travaux que nous présentons dans cette thèse concernent la modélisation des aspects socioculturels et temporels pour permettre aux communautés sénégalaises de partager et de co-construire leur connaissances socioculturelles. En effet, avec la mondialisation la nouvelle génération africaine et particulièrement sénégalaise a de moins en moins de connaissances sur les aspects socioculturels de leur environnement. Ainsi pour rafraîchir la mémoire de nos concitoyens et redonner vie aux nombreux récits qui accompagnent la création et la vie au quotidien des différents terroirs sénégalais, | ||
+ | |||
+ | Nos propositions s' | ||
+ | |||
+ | Nos contributions consistent à (i) proposer des ontologies pour annoter des ressources socioculturelles et (ii) proposer un cadre permettant aux communautés de partager et de co-construire leur connaissances. Les ontologies représentent le socle du Web sémantique et permettent de caractériser un domaine en décrivant les concepts fondamentaux et les relations entre eux. Ainsi, nous avons défini deux ontologies : 1) une ontologie socioculturelle reposant sur la théorie historico-culturelle de l' | ||
====== 2015 ====== | ====== 2015 ====== | ||
===== Soutenances ===== | ===== Soutenances ===== | ||
+ | |||
+ | |||
+ | ==== PhD Thesis Defense - NGUYEN Thi Hoa Hue ==== | ||
+ | |||
+ | '' | ||
+ | |||
+ | == Jury == | ||
+ | Directeur: Monsieur Nhan LE-THANH, Université Nice Sophia Antipolis\\ | ||
+ | Examinateurs: | ||
+ | Rapporteurs: | ||
+ | * Madame GHODOUS Parisa, Université de Lyon I\\ | ||
+ | * Monsieur BELLATRECHE Ladjel, ISAE-ENSMA | ||
+ | |||
+ | |||
+ | == Abstract == | ||
+ | **Title: Control flow-based business workflow templates checking: An | ||
+ | Approach using the Knowledge-based Systems** | ||
+ | |||
+ | This thesis tackles the problem of modelling semantically rich business | ||
+ | workflow templates and proposes a process for developing workflow | ||
+ | templates. The objective of the thesis is to transform a business process | ||
+ | into a control flow-based business workflow template that guarantees | ||
+ | syntactic and semantic validity. The main challenges are: (i) to define a | ||
+ | formalism for representing business processes; (ii) to establish automatic | ||
+ | control mechanisms to ensure the correctness of a business workflow | ||
+ | template based on a formal model and a set of semantic constraints; | ||
+ | (iii) to organize the knowledge base of workflow templates for a workflow | ||
+ | development process. | ||
+ | |||
+ | We propose a formalism which combines control flow (based on Coloured | ||
+ | Petri Nets (CPNs)) with semantic constraints to represent business | ||
+ | processes. The advantage of this formalism is that it allows not only | ||
+ | syntactic checks based on the model of CPNs, but also semantic checks | ||
+ | based on Semantic Web technologies. | ||
+ | |||
+ | We start by designing an OWL ontology called the CPN ontology to represent | ||
+ | the concepts of CPN-based business workflow templates. The design phase is | ||
+ | followed by a thorough study of the properties of these templates in order | ||
+ | to transform them into a set of axioms for the CPN ontology. In this | ||
+ | formalism, a business process is syntactically transformed into an | ||
+ | instance of the CPN ontology. | ||
+ | process becomes simply a verification by inference, by concepts and by | ||
+ | axioms of the CPN ontology on the corresponding instance. | ||
+ | |||
+ | We also introduce the formal definition of semantic constraints, | ||
+ | express dependencies between the activities of a business process. We | ||
+ | present an algorithm to check redundant and conflicting semantic | ||
+ | constraints. | ||
+ | into an instance of a business process ontology called the BP ontology. A | ||
+ | business workflow template is then developed by creating correspondences | ||
+ | between the BP ontology and the CPN ontology. This enables semantic checks | ||
+ | related to a specific business process. | ||
+ | |||
+ | We represent the set of axioms of the CPN ontology related to syntactic | ||
+ | checks as well as the semantic verification issues related to semantic | ||
+ | checks as SPARQL queries. In order to verify workflow templates, we use | ||
+ | the Jena semantic engine to match an RDF graph representing a business | ||
+ | workflow template to graph patterns of these SPARQL queries. If there are | ||
+ | no matches, i.e., no shortcomings, | ||
+ | knowledge base. | ||
+ | |||
+ | In addition, to provide additional support for specifying business rules, | ||
+ | we introduce Event Condition Action (ECA)-like rules that express business | ||
+ | level correctness requirements. The sets of ECA-like rules are stored | ||
+ | along with the corresponding business workflow template in the same | ||
+ | knowledge base. The knowledge base is organized to facilitate the | ||
+ | shareability and reusability of workflow templates. Finally, a prototype | ||
+ | is developed to demonstrate the feasibility and benefits of the approach. | ||
+ | |||
+ | |||
+ | == Résumé == | ||
+ | |||
+ | **Titre: Techniques d' | ||
+ | |||
+ | Cette thèse traite le problème de la modélisation des patrons de workflow | ||
+ | semantiquement riche et propose un processus pour développer des patrons | ||
+ | de workflow. | ||
+ | patron de workflow métier basé sur les flux de contrôle qui garantit la | ||
+ | vérification syntaxique et sémantique. Les défis majeurs sont : (i) de | ||
+ | définir un formalisme permettant de représenter les processus métiers; | ||
+ | (ii) d' | ||
+ | conformité des patrons de workflow métier basés sur un modèle formel et | ||
+ | un ensemble de contraintes sémantiques; | ||
+ | patrons de workflow métier pour le développement de patrons de workflow. | ||
+ | |||
+ | Nous proposons un formalisme qui combine les flux de contrôle (basés sur | ||
+ | les Réseaux de Petri Colorés (CPNs)) avec des contraintes sémantiques pour | ||
+ | représenter les processus métiers. L' | ||
+ | permet de vérifier non seulement la conformité syntaxique basée sur le | ||
+ | modèle de CPNs mais aussi la conformité sémantique basée sur les | ||
+ | technologies du Web sémantique. | ||
+ | |||
+ | Nous commençons par une phase de conception d'une ontologie OWL appelée | ||
+ | l’ontologie CPN pour représenter les concepts de patrons de workflow | ||
+ | métier basés sur CPN. La phase de conception est suivie par une étude | ||
+ | approfondie des propriétés de ces patrons pour les transformer en un | ||
+ | ensemble d' | ||
+ | processus métier est syntaxiquement transformé en une instance de | ||
+ | l’ontologie. La vérification syntaxique d'un processus métier devient | ||
+ | simplement une vérification par inférence, par concepts et par axiomes de | ||
+ | l' | ||
+ | |||
+ | Nous introduisons aussi la définition formelle de contraintes sémantiques, | ||
+ | qui exprime les dépendances entre les activités d'un processus | ||
+ | Nous présentons un algorithme pour la vérification des contraintes | ||
+ | sémantiques redondantes et conflictuelles. Un ensemble de contraintes | ||
+ | sémantiques vérifiées est transformé en une instance de l’ontologie de | ||
+ | processus métier appelée BP-ontology. Un patron de workflow métier est | ||
+ | ensuite développé en créant des correspondances entre l’ontologie BP et | ||
+ | l’ontologie CPN. Il permet les vérifications sémantiques d’un processus | ||
+ | métier spécifique. | ||
+ | |||
+ | Nous représentons l' | ||
+ | conformité syntaxique ainsi que les questions de vérification sémantique | ||
+ | liées à la conformité sémantique en utilisant des requêtes SPARQL. Afin de | ||
+ | vérifier les patrons de workflow, nous utilisons le moteur sémantique Jena | ||
+ | pour l' | ||
+ | métier de ces requêtes SPARQL. Si un patron de workflow métier est | ||
+ | vérifié, il sera stocké dans une base de connaissances. | ||
+ | |||
+ | De plus, dans l’objectif de fournir un soutien supplémentaire pour la | ||
+ | définition de règles métiers, nous introduisons des règles sous forme de | ||
+ | Condition Action Événement (CEA), qui expriment | ||
+ | processus au niveau métier. Les ensembles de règles CEA sont stockés avec | ||
+ | le patron de workflow métier correspondant dans la même base de | ||
+ | connaissances. La base est organisée pour faciliter la capacité de partage | ||
+ | et de réutilisation des patrons de workflow. Enfin, un prototype est conçu | ||
+ | pour démontrer la faisabilité et les avantages de l' | ||
==== PhD Thesis Defense - Somsack INTHASONE ==== | ==== PhD Thesis Defense - Somsack INTHASONE ==== | ||
Line 91: | Line 410: | ||
===== Séminaires ===== | ===== Séminaires ===== | ||
+ | |||
+ | ==== Valerio Basile: Sentiment Analysis on Twitter: The Case of Italian Language ==== | ||
+ | When: 28th Aug. 2015, 11h-12h \\ | ||
+ | Where: EURECOM, room Fourier\\ | ||
+ | By: Valerio Basile | ||
+ | |||
+ | **Abstract** | ||
+ | Sentiment Analysis, also referred to as Opinion Mining, is | ||
+ | the branch of Natural Language Processing that deals with the | ||
+ | identification of personal opinions and emotions in natural language. | ||
+ | With the rise in popularity of Web-based social media platform such as | ||
+ | Facebook and Twitter, the dimension of user-generated content has grown | ||
+ | exponentially, | ||
+ | the researcher as well as to industries and institutions interested in | ||
+ | monitoring the public opinion. | ||
+ | |||
+ | In this talk I will present the work done on the collection of a large | ||
+ | quantity of tweets written in Italian, and the result of the first | ||
+ | experiments of sentiment analysis on such tweets. In the second part, I | ||
+ | will talk about the experience of the Sentiment Polarity Classification | ||
+ | shared task, the most popular task at the EVALITA evaluation campaign of | ||
+ | 2014. The results of the participant systems show how the analysis of | ||
+ | subjective, sometimes ironic, opinions on social media is far from being | ||
+ | a solved problem. Finally, I will spend a few words on what could be the | ||
+ | future of sentiment analysis on social media and how it could | ||
+ | beneficially interact with related areas of natural language analysis | ||
+ | such as Entity Linking. | ||
+ | |||
+ | ** About the speaker**\\ | ||
+ | Valerio Basile is a computer scientist turned to Natural Language | ||
+ | Processing, with particular focus on semantics. Up to June 2015 he was | ||
+ | at the University of Groningen where he obtained his PhD and helped | ||
+ | building the Groningen Meaning Bank, a large collection of semantically | ||
+ | annotated text, and Wordrobe, a Game With A Purpose to collect | ||
+ | linguistic knowledge from the crowd. He is also involved in social media | ||
+ | analysis with TWITA, the collection of Italian tweets, and in Sentiment | ||
+ | Analysis, with the organization of the EVALITA shared task on polarity | ||
+ | classification of Italian tweets (SENTIPOLC). He recently joined the | ||
+ | Wimmics team at INRIA Sophia Antipolis to work on Knowledge | ||
+ | Representation and Information Extraction. | ||
+ | |||
==== Seminary by Claudia d' | ==== Seminary by Claudia d' |