Articles

The Treebanks as the sets of syntactically an- notated sentences, are the most widely used language resource in the application of Natural Language Processing. The occurrence of er- rors in the automatically created Treebanks is one of the main obstacles limiting the using of these resources in the real world …...

This paper aims at effective use of training da- ta by extracting sentences from large general- domain corpora to adapt statistical machine translation systems to domain-specific data. We regard this task as a problem of filtering training sentences with respect to the target domain1 via different similarity metrics. Thus, we …...

Dictionaries are reference resources for learning and diffusing natural languages. Their contents must be enriched carefully due to their impor- tance. However, such contents might contain er- rors and inconsistencies that are hard to detect manually. Several researches have been made in recent years in order to perform this step …...

The most-frequent-sense and the pre- dominant domain sense play an impor- tant role in the debate on word-sense- disambiguation. This discussion is, how- ever, biased by the way sense-tagged cor- pora are built. In this paper, we ar- gue that current sense-tagged corpora ne- glect rare senses and contexts and, …...

In this paper, we introduce a corpus of human-authored dialogue summaries col- lected through a web-experiment. The corpus features (i) one of the few ex- isting corpora of written dialogue sum- maries; (ii) the only corpus available for dialogue summaries in Portuguese; and (iii) the only available corpus of sum- …...

This article reports on mass experiments sup- porting the idea that data extracted from strongly comparable corpora may successfully be used to build statistical machine translation systems of reasonable translation quality for in-domain new texts. The experiments were performed for three language pairs: Spanish- English, German-English and Romanian- English, based …...

The performance of NLP classifiers largely depends on the quality of the features considered for prediction (feature engineering). However, as the number of features increases, the more likely overfit- ting becomes and performance decreases. Also, due to the very large number of features, only slimple linear classifiers are considered, thus …...

This paper provides an analysis of character-level machine translation mod- els used in pivot-based translation when applied to sparse and noisy datasets, such as crowdsourced movie subtitles. In our experiments, we find that such character- level models cut the number of untrans- lated words by over 40% and are espe- …...

Sublanguages are specialized genres of language associated with specific domains and document types. When sublanguages can be recognized and adequately charac- terized, they are useful for a variety of types of natural language processing ap- plications. Although there are sublan- guage studies related to languages other than English, all previous …...

We describe in this paper how different learning strategies can be applied on the same NLP task, namely chunking. The reference corpus is extracted from the French Treebank, the symbolic learning strategy used is grammatical inference and the statistical one is CRFs (Conditional Random Fields). As expected, the sym- bolic …...

Analyzing how people discuss about health-related topics on dedicated forums and social networks such as Twitter, can provide valuable insight for syndromic surveillance and to predict disease outbreaks. In this paper we present a minimally trained algorithm to learn associations between technical and everyday language terms, based on pattern generalization …...

With 19%–28% of Internet users partici- pating in online health discussions, it be- came imperative to be able to detect and analyze posted personal health informa- tion (PHI). In this work we introduce two semantic-based methods for mining PHI on social networks which will warn the users about potential privacy …...

In this work we present sentiment analysis of messag- es posted on a medical forum. We categorize posts, written in English, into five categories: encourage- ment, gratitude, confusion, facts, and facts + senti- ments. Our study applies a manual sentiment annota- tion, affective lexicons in its sentiment analysis and machine …...

The purpose of part-of-speech tagging is to automatically tag the words of a text, written in a certain language, with labels that usually take the form of acronyms that designate the appropriate parts-of-speech. In this paper we propose a new approach to the problem that divides it in two different …...

We report on the recent development of ParZu, a German dependency parser. We discuss the effect of POS tagging and morphological analysis on parsing perfor- mance, and present novel ways of improv- ing performance of the components, in- cluding the use of morphological features for POS-tagging, the use of syntactic …...

Patent search is an important information retrieval problem in scientific and busi- ness research. Semantic search would be a large improvement to current technolo- gies, but requires some insight into the lan- guage of patents. In this article we test the fit of the language of patents to the sublan- …...

This paper proposes a two-step approach to find hypernym relations between pairs of noun phrases in Dutch text. We first apply a pattern-based approach that com- bines lexical and shallow syntactic infor- mation to extract a list of candidate hy- pernym pairs from the input text. In a second step, …...

We present an unsupervised learning model that induces phrasal inversion transduction grammars by introducing a minimum conditional description length (CDL) principle to drive search over a space defined by two opposing extreme types of ITGs. Our approach attacks the difficulty of acquiring more complex longer rules when inducing inversion transduction …...

In this paper the author presents a new context independent method for bilingual term map- ping using maximised character alignment maps. The method tries to particularly address mapping of multi-word terms and compound terms that are extracted from comparable cor- pora. The method allows integrating linguistic resources (e.g., probabilistic dictionaries …...

The lack of labeled data always poses challenges for tasks where machine learn- ing is involved. Semi-supervised and cross-domain approaches represent the most common ways to overcome this dif- ficulty. Graph-based algorithms have been widely studied during the last decade and have proved to be very effective at solv- ing …...