What is “AI, Cognitive Semantics and Computational Linguistics : New Perspectives” ?

Traditionally, the study of computational linguistics has been performed by computer scientists, specializing in the application of computers to the processing of a natural language.

Today, computational linguists often work as members of interdisciplinary teams, including linguists (specifically trained in linguistics), language experts (persons with some level of ability in the languages relevant to a given project), and computer scientists. In general, computational linguistics draws upon the involvement of linguists, computer scientists, experts in artificial intelligence, mathematicians, logicians, cognitive scientists, cognitive psychologists, psycholinguists, anthropologists and neuroscientists, among others.

Computational linguistics must become more connected to the cognitive sciences through the development of cognitive semantic theories. computational linguistics is connected to artificial intelligence through the development of methods and algorithms for all aspects of language analysis and their computer implementation. We can see language analysis divided into two parts: theoretical analysis and application. The theoretical aspect includes standard areas studied in linguistics: semantics, syntax, and morphology. Semantic theories guide the development of syntactic theories and morphology. Semantic theories can be based on some specific features of computation, but at the present stage of research, there is a gap between linguistic analysis and computer applications in two senses: there are many computer applications without linguistic theoretical support and, conversely, there are a number of theoretical methods with no computer implementation. Another epistemological feature of the present stage of research is that most computational linguistic methods are focused on statistical approaches. The advantage of these methods is that they are easy to apply but the drawback is that they distort the qualitative and genuine cognitive features of language.

The goal of this track is to provide an international forum for discussing the latest approaches in subfields of computational linguistics related to cognitive semantics and to artificial intelligence. Its aim is also to exchange ideas concerning the way of building efficient systems of language analysis based on cognitive semantic models.

Its mission is to prove the increase of interaction between modeling in cognitive semantics and computer implementation: a good system analysis requires a good semantic model as framework. On the other hand, the need for semantic models other than those based on statistical methods has to be proved.


What is the GOAL of the track?

This track is intended to present works ranking from logical and mathematical models in syntax and semantics (logic of objects, topological theories of time and space etc.) as foundations of the design and analysis to natural language processing systems.


Who should be interested?

Special tracks, held in parallel with the general conference, are an integral part of the conference. They provide researchers in focused areas the opportunity to meet and present their work, and offer a forum for interaction among the broader community of artificial intelligence researchers.

Topics of interest are in all areas related to artificial intelligence. For example, last year's special tracks included: AI Education; AI Planning and Scheduling; Applied Natural Language Processing; Case-Based Reasoning; Data Mining; Design, Evaluation, and Refinement of Intelligent Systems (DERIS); Games and Entertainment; Intelligent Tutoring Systems; Semantics, Ontologies, and Computational Linguistics; and Uncertain Reasoning.


What kind of studies will be of interest?

Papers and contributions are encouraged for any work relating to AI, Cognitive Semantics and Computational Linguistics. Topics of interest may include (but are in no way limited to)

  1. cognitive semantics,

  2. logics of language,

  3. language modeling,

  4. computational linguistics (lexicology; morphology; syntax; semantic).



Any paper related to cognitive semantic and computational linguistics is welcome.


Note: We invite original papers (i.e. work not previously submitted, in submission, or to be submitted to another conference during the reviewing process).

All accepted papers will be published as FLAIRS proceedings by the AAAI.


Other tracks at FLAIRS that may be of interest to those submitting to this track include Cognition and AI: Capturing Cognitive Plausibility and Informing Psychological Processes, Applied Natural Language Processing (http://tinyurl.com/kwwzs8) in a total of 14 special tracks that accompany the main conference.

 

Schedule

Wednesday, 19th May

Chair: Anca Pascu

1:45pm Annotating Lexically Entailed Subevents for Textual Inference Tasks Seohyun Im,
James Pustejovsky

2:10pm French-written Event Extraction Based on Contextual Exploration
Aymen Elkhlifi, Rim Faiz
Event extraction is a significant task in information extraction. This importance increases more and more with the explosion of textual data available on the Web, the appearance of Web 2.0 and the tendency towards the Semantic Web. Thus, we propose a generic approach to extract French-written events from text and analysis them. We propose an event extraction algorithm with a polynomial complexity O(n5), and a new similarity measurement between events. We use this measurement to gather similar events. We also present a semantic cart of events, and we validate the first component of our approach by the development of the "EventEC" system.

2:35pm Direct Reported Speech in Multilingual Texts: Automatic Annotation and Semantic Categorization
Motasem Alrahabi, Jean-Pierre Desclés, Jungyeon Suh
We propose an application for the automatic identification and categorization of quotations. The categorization is based on a semantic map of enunciative modalities. The texts are treated in three languages: Arabic, Korean and French.

 

Chair: Ismail Biskri

3:30pm Invited Talk: Reasoning in Natural Language in using Combinatory Logic and Topology
Jean-Pierre Desclés, Universite de Paris-Sorbonne, France
Curry's Combinatory Logic (CL) is a logical formalism to study the intrinsic compositions of operators. Aspect is an operator 'ASPI' applied to a predicative relation realized onto a topological interval 'I' of instants (that is: true at each instant of this interval 'I'). Different aspectual values of 'ASPI' are concerned: state (STATE) or process (PROC) or event (EVEN) realized onto different types of topological intervals. It is shown how CL is used to take into account: (i) formal representations of verbal meanings by cognitive schemes; (ii) aspectual values and temporal relations between topological intervals in a computational and cognitive approach of the relations to speaking act. An example of a "natural inference" between two utterances like John took Mary's pen → Now, John has got the pen is used to explain the different steps of a formal processing with the framework of a polystratal computational model - Cognitive Applicative Grammar - which provides an useful interplay between cognitive representations and morpho-syntactic configurations by means of compositions of operators by combinators of the CL and functional types of Categorial Grammars.

Jean-Pierre Desclés is professor of Computer Science and Linguistics at the Sorbonne; he has been the head of the Graduate School « Concepts and Languages » and he is the director of the laboratory LaLIC ("Languages, Logics, Informatics and Computer Sciences") at the University of Paris-Sorbonne. He has studied the mathematics, computer sciences and linguistics at Paris. He is a member of the International Academy of Philosophy of Sciences (AIPS, Bruxelles). He has worked, as mathematician and computer scientist, with the linguists S.K. Schaumyan (Yale University), A. Culioli, Z. Guentchéva (Paris). He has worked in different fields : (i) theoretical linguistics (aspects and tenses, spatial representations, diatheses); (ii) Combinatory Logics and λ-calculus with types applied to an analysis of semantic problems by means of a computational model - Applicative and Cognitive Grammar (ACG) - and also, with Anca Pascu, to a logical analysis of concepts, with typical and atypical instances, in the framework of LDO - Logics of Determination of Objects - ; (iii) interactions between cognitive semantics and computational representations of the meaning of verbs and prepositions for building formal ontologies; (iv) automatic annotations of texts from discursive and semantic viewpoints by means of a contextual exploration strategy. He is author of several books and articles on relations between cognition, computation and semantic analysis of natural languages.

 

Thursday, 20th May

Chair: Susan Haller

10:30am Disambiguation of Textual Data Typification for the Purpose of Categorial Analysis
Adam Joly, Ismail Biskri, Boubakar Hamrouni
The operation of categorial type assignment is prior to a categorial analysis. The lexical units, which are entries in a dictionary, are commonly associated to one or more categorial types. Therefore, we need to determine for each unit in a sentence the correct categorial type to be assigned. Current research on Categorial Grammars is not paying attention enough to this issue. The assignment of categorial types is often done either manually, or with ad hoc heuristics. In this paper we present a method based on conditional probabilities.

10:55am Coordination of Standard Arabic Subject Markers: Implementing the Agreement Asymmetries in the ACCG Framework
Ismaïl Biskri, Louisette Emirkanian, Adel Jebali
In Standard Arabic, object markers and subject markers behave differently, although they share some properties. We are concerned here by their morphosyntactic status; whether they are arguments or agreement markers. The status of object markers is not an issue by itself, but the status of subject markers is one. The agreement asymmetries lead us to stipulate that, when subject markers are doubled by NPs whether these NPs are coordinated or not, they are agreement markers and not arguments. This analysis is implemented by means of an applicative combinatory categorical grammar (ACCG).

11:20am Explanation Versus Meta-Explanation: What Makes a Case More Convincing?
Boris Galitsky, Josep Lluis de la Rosa, Boris Kovalerchuk
Comparative analysis of the roles of explanation and meta-explanation is conducted assessing the validity of explanation exchanged between human agents. Meta-explanation links the overall structure of a current scenario with that of previously learned scenarios of multi-agent interaction. The scenario structure includes communicative actions of involved agents and argumentation attack relations between the subjects of these actions. Object-level explanation is based on a traditional machinery to handle argumentative structure of a dialogue, assessing the plausibility of individual claims. To assess plausibility of customer complaints, we relate them to the classes of valid (consis-tent, genuine) and invalid (inconsistent, include attempts to get compensation from a company, or expressing a bad mood). Evaluation of contribution of each argumentation level shows that both levels of explanation are essential for assessment of whether a multi-agent scenario as described by an agent is plausible or not. We demonstrate that meta-explanation in the form of machine learning of scenario structure should be augmented by conventional explanation by finding factual-based arguments for individual claims. We also define a ratio between object-level and meta-explanation as relative accuracy of plausibility assessment based on former and latter sources. We then observe that groups of scenarios can be characterized based on a specific ratio between object-level and meta-level explanations in a phase space; such ratio is an important parameter of human behavior associ-ated with explaining in a dialogue.

 

Chair: Florence Le Priol

1:45pm Inverting Semantic Structure under Open Domain Opinion Mining
Boris Galitsky, Josep Lluis de la Rosa, Gábor Dobrocsi
We explore the semantic structure of how opinions on products and services are ex-pressed in blogs and forums in the form of user needs. To optimize the efficiency of con-tent delivery, we invert the product-feature structure and propose a specific way to repre-sent the user opinion content in forums and blogs, focusing on user needs about product qualities and features. The content is subject to inversion so that these needs become pri-mary entry points for browsing and search. User need is defined syntactically; semantic and concept structure means for such user needs are developed. The system is subject to evaluation with respect to coverage and in-formation access efficiency.

2:10pm Combining MT Systems Effectively
Petr Homola, Jernej Vicic
The paper describes a sophisticated method of combining two MT systems to obtain a new translation pair. Instead of a simple pipe, we use a complex data structure to pass the data from the first MT system to the second one. Evaluation results are reported for the language triplet Czech-Slovenian-Slovak.

 

Co-chairs :

Ismail Biskri, Universite de Québec à Trois Rivières, Canada

Susan Haller, State University of New York at Potsdam, USA

Florence Le Priol, Université de Paris-Sorbonne, Paris, France

Anca Pascu, Université de Bretagne Occidentale, Brest, France

James Pustejovsky, Brandeis University, USA


Program commitee :

Maryvonne Abraham, Institut Telecom, Telecom-Bretagne, France
David Banks, Université de Bretagne Occidentale, Brest, France
Ismail Biskri (co-chair), Universite de Québec à Trois Rivières, Canada,
Alex Borgida, Rutgers University, USA
Walter Carnielli, University of Campinas, Brazil
François-Gilles Carpentier, Université de Bretagne Occidentale, Brest, France
Fintan Costello, University College Dublin, Ireland
Dan Cristea, University of Iasi, Romania
Richard Dapoigny, Université de Savoie, France
Jean-Pierre Desclés, Université Paris-Sorbonne, France
Michel De Glas, Ecole Polytechnique, France
Kathelijne Denturck, Ecole de Traduction et d'Interprétation, Haute Ecole de Gand, Belgium
Brahim Djioua, Université Paris-Sorbonne, France
Rim Faiz, IHEC de Carthage, Université du 7 novembre à Carthage, Tunisia
Boris Galitsky, Knowledge Trail, MA, USA
Vera Goodacre, George Mason University, USA
Zlatka Guentchéva, CNRS, France
Ewa Gwiazdecka, Universiy of Warsawa, Poland
Susan Haller (co-chair), University of Wisconsin, USA
Eva Hajiova, Charles University, Prague, Czech Republic
Adel Jebali, University Concordia, Montreal, Canada,
Klara Ladji, University of Tirana, Albania
Guy Lapalme, Université de Montréal, Canada
Peter Lazarov, Université de Sofia, Bulgaria
Florence Le Priol (co-chair), Université Paris-Sorbonne, France
Jean-Guy Meunier, Université de Québec à Montréal,Canada
Ghassan Mourad, Université de Beirouth, Lebanon
Anca Pascu (co-chair), Université de Bretagne Occidentale, Brest, France
Patrice Pognan, INALCO, Paris, France
James Pustejovsky (co-chair), Brandeis University, USA
Gilles Richard, British Institute of Technology & E-commerce, London, UK
Christophe Roche, Université de Savoie, France
Benoît Sauzay, France Telecom
Jong-Seok Soh, Hankuk University of Foreign Studies, Seoul, Korea
Jungyeon Suh, Seoul Women's University, Seoul, Korea
Jana Sukkariek, ETS, USA
Geoffrey Williams, Université de Bretagne Sud, Vannes, France

 

Dates and Deadlines

Paper submission deadline November 23rd, 2009
Notification of paper decisions January 22nd, 2010
Camera ready version due February 22, 2010
Conference 19-May-2010 - 21-May-2010

 

Submission Instructions

Interested authors (for the general track or the special topic) should format their papers according to AAAI formatting guidelines. The papers should be original work (i.e., not submitted, in submission, or submitted to another conference while in review). Papers should not exceed 6 pages (2 pages for a poster) and are due by November 23rd, 2009. For FLAIRS-23, the 2010 conference, the reviewing is a double blind process. Fake author names and affiliations must be used on submitted papers to provide double-blind reviewing. Papers must be submitted as PDF through the EasyChair conference system, which can be accessed through the main conference web site (http://www.FLAIRS-23.info).

Note: do not use a fake name for your EasyChair login - your EasyChair account information is hidden from reviewers. Authors should indicate the special track for submissions. The proceedings of FLAIRS will be published by the AAAI. Authors of accepted papers will be required to sign a form transferring copyright of their contribution to AAAI. An author of each accepted paper is required to register, attend, and present the paper at FLAIRS.

Please, check the website http://www.flairs-23.info/ for further information.