Articles

Tonal errors pose a serious problem to Mandarin Chinese learners, making them stumble in their communication. The purpose of this paper is to investigate beginner level Japanese students’ difficulties in the perception and pronunciation of disyllabic words, particularly to find out which combinations of tones these errors mostly occur in. …...

We investigate whether suffix related features can significantly improve the performance of character-based approaches for Chinese word segmentation (CWS). Since suffixes are quite productive in forming new words, and OOV is the main error source for CWS, many researchers expect that suffix information can further improve the performance. With this …...

With the popularity of network application- s, new words become more common and bring the poor performance of natural lan- guage processing related applications in- cluding web search. Identifying new words automatically from texts is still a very chal- lenging problem, especially for Chinese. In this paper, we propose a …...

Towards a treebank of spoken French We present the first results of an attempt to build a spoken treebank for French. It has been conducted as part of the ANR project Etape (resp. G. Gravier). Contrary to other languages such as English (see the Switchboard treebank (Meteer, 1995)), there is …...

s Pin-ning Tu Shih-Ping Wang National Taiwan University of Science and Technology, Taiwan/ #43, Sec. 4, Keelung Rd. Da’an Dist., Taipei 106, Taiwan (R.O.C.) National Taiwan University of Science and Technology, Taiwan/ #43, Sec. 4, Keelung Rd., Da’an Dist., Taipei 106, Taiwan (R.O.C.) promise523@hotmail.com spwang2005@yahoo.com.tw Abstract There has long been …...

Studies of spontaneous conversational speech grounded on large and richly annotated cor- pora are still rare due to the scarcity of such resources. Comparative studies based on such resources are even more rarely found because of the extra-need of comparability in terms of content, genre and speaking style. The present …...

The demand is increasing recently for non- task-oriented conversation system in various scenes. Previous studies provide various strategies to enrich the methods for generating utterances, thus making the conversation systems or agents appear more interesting. However, most previous works tend to rely on templates and therefore are not able to …...

Generation System for Social Scientific Papers Michio Kaneko Dongli Han Graduate School of Integrated Basic Sciences Department of Information Science, Nihon University, Tokyo, JAPAN College of Humanities and Sciences m-kaneko@chs.nihon-u.ac.jp Nihon University, Tokyo, JAPAN han@chs.nihon-u.ac.jp Abstract Abstracts are quite useful when one is trying to understand the content of a …...

Head-Internal Relatives (HIRs) in Japanese are regarded as rich context-setters within Dynamic Syntax (DS): the propositional tree of the HIR clause is mapped onto a ‘partial’ tree, which establishes a rich context for the embedding clause to be parsed. This partial tree contains a situation node decorated with the Relevancy …...

The use of an obligatory numeral classifier (C) on N in general does not co-occur with mandatory plural marking (PM) (Greenberg 1990[1972], Sanches and Slobin 1973). Borer (2005) and Her (2012a) take this generalization further and see Cs and PMs as the same category. This unification implies that C/PM are …...

It has been disputed whether scalar implica- tures (= SIs) arise globally or locally. Basi- cally SIs should be global because they arise by comparing strengths of whole alternative statements. On the other hand, there are a lot of examples in which local SIs are prefer- able. Linguists like Chierchia …...

_______________________________________________________________________ Automatic tagging of a learner corpus of English with a modified version of the Penn Treebank tagset This article covers the issue of automatic annotation of a learner corpus of English. The objective is to show that it is possible to PoS-tag the corpus with a tagger to prepare …...

Unsupervised CRF for knowledge discovery Knowledge discovery aims at bringing out coherent groups of entities. They are usually based on clustering ; the challenge is then to define a notion of similarity between the relevant entities. In this paper, we propose to divert Conditional Random Fields (CRF), which have shown …...

Construction of a Free Large Part-of-Speech Annotated Corpus in French This paper studies the possibility of creating a new part-of-speech annotated corpus in French from an existing one. The objectives are to propose an exit from the restrictive licence of the source corpus and to obtain a perpetual modernisation of …...

A Supervised learning for the identification of semantic relations in parallel enumerative structures This work falls within the framework of ontology engineering and learning from encyclopedic or scientific texts. Our original contribution lies within the extraction of semantic relations expressed beyond the text linearity. To this end, we relied on …...

Learning a hierarchy of specialized pairwise models for coreference resolution This paper proposes a new method for significantly improving the performance of pairwise coreference models. Given a set of indicators, our method learns how to best separate types of mention pairs into equivalence classes for which we construct distinct classification …...

Discriminative statistical approaches for multilingual speech understanding Statistical approaches are now widespread in the various applications of natural language processing and the elicitation of an approach usually depends on the targeted task. This paper presents a comparison between the methods used for machine translation and speech understanding. This comparison allows …...

Automatically identifying implicit discourse relations using annotated data and raw cor- pora This paper presents a system for identifying « implicit » discourse relations (that is, relations that are not marked by a discourse connective). Given the little amount of available annotated data for this task, our system also resorts …...

Grouping of terms based on linguistic and semantic regularities in a cross-lingual context We propose to exploit the Natural Language Processing methods dedicated to terminology structuring independently in two languages (English and French) and then to merge the results obtained in each language. The terms are grouped into clusters thanks …...

Unsupervised selection of semantic relations for improving a distributional thesaurus Work about distributional thesauri has shown that the relations in these thesauri are mainly reliable for high frequency words. In this article, we propose a method for improving such a thesaurus through its re-balancing in favor of low frequency words. …...