Dan Cristea

Veins Theory: A Model of Global Discourse Cohesion and Coherence

In this paper, we propose a generalization of Centering Theory (CT) (Grosz, Joshi, Weinstein (1995)) called Veins Theory (VT), which extends the applicability of centering rules from local to global discourse. A key facet of the theory involves the identification of «veins» over discourse structure trees such as those defined in RST, which delimit domains of referential accessibility for each unit in a discourse. Once identified, reference chains can be extended across segment boundaries, thus enabling the application of CT over the entire discourse. We describe the processes by which veins are defined over discourse structure trees and how CT can be applied to global discourse by using these chains. We also define a discourse «smoothness» index which can be used to compare different discourse structures and interpretations, and show how VT can be used to abstract a span of text in the context of the whole discourse. Finally, we validate our theory by analyzing examples from corpora of English, French, and Romanian.

Expectations in incremental discourse processing

The way in which discourse features express connections back to the previous discourse has been described in the literature in terms of adjoining at the right frontier of discourse structure. But this does not allow for discourse features that express expectations about what is to come in the subsequent discourse. After characterizing these expectations and their distribution in text, we show how an approach that makes use of substitution as well as adjoining on a suitably defined right frontier, can be used to both process expectations and constrain discouse processing in general.

Summarisation through discourse structure

In this paper we describe a method to obtain summaries focussed on specific characters of a free text. Summaries are extracted from discourse structures which differ from RST structures by the fact that the trees are binary and lack relation names. The discourse tree structures are obtained by combining constraints given by cue-phrases (resembling Marcus method) with constraints coming from the exploitation of cohesion and coherence properties of the discourse (as proved by Veins Theory). The architecture of a summarisation system is presented on which evaluations intended to evidence the contribution of each module in the final result are performed and discussed.

Improving the search for learning objects with keywords and ontologies

We report on an ongoing project which aims at improving the effectiveness of retrieval and accessibility of learning object within learning management systems and learning object repositories. The project Language Technology for eLearning approaches this task by providing Language Technology based functionalities and by integrating semantic knowledge through domain-specific ontologies. We will report about the development of a keyword extractor and a domain-specific ontology, the integration of these modules into the learning management system ILIAS and the validation of these tools which assesses their added value in the scenario of searching learning objects across different languages.

An empirical investigation of the relation between discourse structure and co-reference

We compare the potential of two classes of linear and hierarchical models of discourse to determine co-reference links and resolve anaphors. The comparison uses a corpus of thirty texts, which were manually annotated for co-reference and discourse structure.

An Incremental Discourse Parser Architecture

We present a discourse parsing architecture based on an incremental approach and aimed at building a rhetorical structure of a free text. Vein expressions computed on the developing structure help to restrict the domains of referential accessibility on which resolution of anaphora is performed. The parsing process is guided by cohesion and coherence constraints.

Summarizing Short Texts Through a Discourse-Centered Approach in a Multilingual Context

The chapter presents the architecture of a system targeting summaries of short texts in six languages. At the core of a summary, which comprises clauses and sentences extracted from the original text, is the structure of the discourse and its relationship with its coreferential links. The approach shows a uniform design for all languages, while language specificity is attributed to the resources that fuel the component modules. The design described here includes a number of feedback loops used to fine-tune the parameters by comparing the output of the modules against annotated corpora. “Average” summaries over some human-produced ones are used to evaluate the accuracy of each of the monolingual systems. The study also presents some quantitative data on the corpora used, showing a comparison among languages and results that, mostly, prove to be above the state of the art.

Multi-Dimensional Analysis of Political Language

This paper presents a method for the valuation of discourses from different linguistic perspectives: lexical, syntactic and semantic. We describe a platform discourse analysis tool (DAT) which integrates a range of language processing tools with the intent to build complex characterisations of the political discourse. The idea behind this construction is that the vocabulary and the clause structure of the sentence betray the speaker’s level of culture, while the semantic classes mentioned in a speech characterises the speaker’s orientation. When the object of study is the political discourse, an investigation on these dimensions could put in evidence features influencing the electing public. The method is intended to help both political speakers to improve their discourse abilities, by comparing their speeches with those of personalities of the public life in the past, and the public at large by evidencing hidden aspects of the linguistic and intellectual abilities of their candidates.

Computational Techniques in Political Language Processing: AnaDiP-2011

This paper presents a computational method, AnaDiP-2011, based on natural language processing (NLP) techniques for the interpretation of the political discourse. The application considers the 2009 presidential campaign in Romania. The concept behind this method is that the manner in which individuals speak and write betrays their sensibility. Our investigation is intended to give support to researchers, specialists in political sciences, political analysts and election’s staff, being helpful mainly in their social exploration of the electoral campaigns in their intend to measure reactions with respect to the developments in the political scene.

Cross-lingual Romanian to english question answering at CLEF 2006

This paper describes the development of a Question Answering (QA) system and its evaluation results in the Romanian-English cross-lingual track organized as part of the CLEF 2006 campaign. The development stages of the cross-lingual Question Answering system are described incrementally throughout the paper, at the same time pinpointing the problems that occurred and the way they were addressed. The system adheres to the classical architecture for QA systems, debuting with question processing followed, after term translation, by information retrieval and answer extraction. Besides the common QA difficulties, the track posed some specific problems, such as the lack of a reliable translation engine from Romanian into English, and the need to evaluate each module individually for a better insight into the systems failures.