Portail TALC

06. I-FLEG

Celine Moro, didiot, Gabrielle Blackburn — 2013-04-11T11:23:51Z

Aidez la recherche, jouez avec IFLEG !

I-FLEG est un jeu 3D qui vise à enseigner le français langue étrangère. Développé par des chercheurs, c'est une première maquette que nous cherchons à améliorer pour illustrer le potentiel de la recherche dans le domaine de l'apprentissage des langues assisté par ordinateur.

En jouant avec IFLEG, vous pratiquez votre français et vous nous aidez dans nos recherches !

L'action se déroule dans une maison, au sein de la communauté en ligne "Second Life". Avec ce jeu, un débutant peut apprendre du vocabulaire, et quelques bases de grammaire, en déplaçant son avatar dans la maison et en cliquant sur les objets qui s'y trouvent. En fonction du défi que vous avez choisi, vous serez soit simplement invité à identifier les objets soit à les qualifier (en faisant attention aux accords et aux déterminants !) etc.

Trois étapes pour jouer avec IFLEG :

Installer un viewer Second Life tel que Phoenix pour MacOSX, Windows et Linux puis créer son avatar
Allez dans l'ilôt IFLEG
Utiliser les notes IFLEG pour vous guider dans le jeu.

Suivez IFLEG dans les réseaux sociaux :

Corpus Linguistics and Methodologies for Human Annotation [LT-M1, optional]

2012-12-14T17:40:52Z

The objective of this course, is to make the students aware of the key role played by annotated textual corpora in contemporary Natural Language Processing (NLP). The importance of methodology and evaluation will be highlighted. - Optional UE

Computational Semantics [CS-M2 or LT-M3, Optional]

2012-12-14T17:40:49Z

This course, which is optional, provides a detailed and mathematically rigorous introduction to semantic construction and the syntax/semantics interface. The underlying mechanisms taught are those of categorial-style grammars (on the syntactic side) and typed lambda calculus (on the semantic side). That is, the fundamental mathematical tools are those pioneered by Montague in his pioneering work form the early 1970s. But the intervening years have seen important mathematical, linguistic, and computational developments in such approaches and this course will introduce the student to contemporary developments.

Grammatical Formalisms [LT-M2 or LT-M3, Optional]

2012-12-14T17:40:45Z

This course, which is optional, aims to introduce the student to some of the grammatical formalisms used in modern computational linguistics. Typically the course will introduce at least two of the following formalisms:

Lexical Functional Grammar
Categorial Grammar
Tree Adjoining Grammar
Interaction Grammar

Just as important as the formalisms themselves are the general concepts they embody, themes such as the importance of features and unification, of the importance of the lexicon, and of campatability with other linguistic levels (such as semantics). After this course the student will be in a position to appreciate the positive and negative points of various approaches.

Lexicology and lexicography [LT-M1, optional]

2012-12-14T17:40:42Z

According to the Oxford English Dictionary, a dictionary is “a book dealing with the individual words of a language (or certain specified classes of them), so as to set forth their orthography, pronunciation, signification, and use, their synonyms, derivation, and history, or at least some of these facts”. This course (an optional course) deals with different elements of this definition: what is a word? What is word signification, and how can it be approached in dictionaries? How does lexical use interact with the system? From where do words come and what is their internal history?

Topics that will be covered include:

A general introduction to the field

Synchronic lexicology and lexicography

Etymology and historical lexicology: inherited lexicon, borrowings, and internal creations

Computerisation of the Trésor de la Langue Française (TLF)

Both traditional and computational lexicography will be addressed. Although most examples will be taken form French and other Romance languages, the lexicon of any given language can come into focus if the students so wish.

Models of Discourse and Dialogue [LT-M3 or CS-M2, Optional]

2012-12-14T17:33:37Z

This course, which is optional, aims to introduce the student to computational models of discourse and/or dialog; in some years the emphasis is on discourse, in others on dialog. Typical topics covered in this course include:

Models for discourse and dialogue
Anaphora resolution and presupposition
Discourse management and grounding
Discourse Representation Theory

However whatever topics are covered, one theme is always present: the importance of inference. Inference is discussed both at the computational level (for example, we discuss and motivate the use of such software tools as planners, model builders, model checkers and theorem provers) and at the theoretical level (for example we discuss the use of inference techniques such as default reasoning and abduction that are typical of pragmatic reasoning).

Claire Gardent — 2012-10-31T10:22:02Z

- SYFRAP

The Emospeech Dialog Corpus

Claire Gardent — 2012-09-24T13:50:06Z

The Emospeech dialog corpus is a corpus of human-machine dialogs in French that were collected in the context of the serious game Mission Plastechnology. This game is designed to promote careers in the plastic industry. It is a multi-player quest where the players (3 teenagers) seek to build a video game joystick in order to free their uncle trapped in the game. To build this joystick, the players must explore the plastic factory and achieve 17 mandatory goals (find the plans, get the appropriate mould, retrieve some plastic from the storing shed, etc).

The collected data consists of 1249 dialogs, 10454 utterances and 168509 words.

Download the Emospeech Corpus : Click Here

Annotation

The annotation schema designed for the game combines core communicative acts (Bunt et al., 2010) with domain specific information. The latter defines the goals being pursued/discussed/achieved etc., while the communicative act can be viewed as specifying how the current system state is updated by the speaker's utterance.

Publications

L. M. Rojas-Barahona, A. Lorenzo and C. Gardent Building and Exploiting a Corpus of Dialog Interactions between French Speaking Virtual and Human Agents . LREC 2012, The eighth international conference on Language Resources and Evaluation, PDF
L. M. Rojas-Barahona, A. Lorenzo and C. Gardent An End-to-End Evaluation of Two Situated Dialog Systems . SIGDIAL 2012, PDF
L. M. Rojas-Barahona and C. Gardent What Should I Do Now? Supporting Progress in a Serious Game . SemDIAL, The 16th Workshop on the Semantics and Pragmatics of Dialogue, Paris, September 19-21, 2012. Demo Paper. PDF

Preparatory Courses (For All Students)

didiot — 2012-09-10T18:51:05Z

In the first three weeks of September we offer a number of preparatory talks and courses. In particular, we offer an intensive course in French (which continues in less intensive form until Christmas) and lectures providing background in programming and computing. During this period you will also discuss possible course options with members of the teaching staff. Following is the arrangement of the preparatory courses.

- Python Introduction (Guy Perrier)

Please refer to the teacher for more information.

- Linux/UNIX & LaTex Introduction (Sai Qian)

This is a 3-days intensive course for the introduction to Linux&Unix/Latex. Slides for each day's course will be distributed after the lecture.

Day 1 (4 September, 2012) - Introduction to Linux/UNIX, operation system, the history and features of Linux/UNIX will be talked briefly first, then the commonly used, also the most basic shell commands for Linux/UNIX will be presented. The lecture will be carried out with the students' practice.

Day 2 (5 September, 2012) - Introduction to Latex, first some basic concepts in Latex and the advantages of doing typesetting with Latex will be introduced. Then the students are required to carry out some practical practice, such as creating and compiling the first latex document (if they have not done that before). Later, the special symbols (such as blank space, newline command) in Latex, the setting for fonts and the heading structure will be described, with the practice of the students again.

Day 3 (6 September, 2012) - More technique details in Latex, based on the knowledge of Day 2, students will go deeper in Latex, such as doing mathematical formulae and creating special typesetting features (such as inserting table, pictures, listing environments) using Latex.

Day 4 (7 September, 2012) - At first, setting up bibliography in Latex will be introduced, which is of great importance when students want to write technique report or term paper or thesis. Further more, students will learn how to create presentation slides with Latex Beamer package, various features will be discussed, such as frame, theme, overlays, and etc.

Go back to the course page

Applications for NLP [LT-M2 and LT-M3, Compulsory 1+2]

2012-09-10T18:44:04Z

This course is compulsory for the EM/LCT students. As its name implies, it aims to introduce the topic of Natural Language Processing (NLP) applications to students. The course has two parts: the lectures and the software project.

1. Lecture (Sai Qian)

The lecture starts with a basic introduction to the motivation and history of NLP, together with current developments in this area. It then presents the relation between linguistics and NLP, using the five conventional evels in linguistics (Phonology/Phonetics, Morphology, Syntax, Semantics and Pragmatics). After that, some more specific NLP applications will be described in details, such as Information Retrieval (IR), Information Extraction (IE) and Classification & Clustering. Finally, the NLTK (Natural Language Toolkit) in Python will be shown to students, who will have the chance to see how the theoretical matter in the first part of the course can really be carried out, without great difficulties.

2. Software Project (Claire Gardent)

The course has a substantial practical component. At the start of the course, software projects will be assigned, and students be grouped into project teams to solve them collaboratively. Every group is required to present the result of their project at the end of the semester. More information on the project will be distributed in the first 2 weeks of the course.

Attention: the course also requires that the students attend the TALC Research Seminar on every Tuesday afternoon, regular attendances will also count in the final result. For the timetable for this, refer to the webpage of the seminar.

Go back to the course page