Related Papers. This work has been presented at COLING 2010.
Related Papers. The SCF system for Italian has been presented as a paper at LREC 2012.
Related Papers. This SCF parser combination work was presented at the LREC Workshop on Language Resource Merging.
Related Papers. Exploration of the lexical characteristics of the biomedical domain can be found in a COLING 2010 paper. • Xxxxxxxx, Xxxx; Xxxxxxxxxx, Xxx; x Xxxxxxxx, Diarmuid; Xxx, Xxx. (2010). Exploring variation across biomedical subdomains . In Xxxxx; Xxx-Xxx and Xxxxxxxx, Xxx (Eds.). Proceedings of the 23rd International Conference on Computational Linguistics (Coling 2010). Beijing, China: Coling 2010. One of the most noticeable trends in the past decade of NLP research has been the deployment of language processing technology to meet the information retrieval and extraction needs of scientists in other disciplines. This meeting of fields has proven mutually beneficial: scientists increasingly rely on automated tools to help them cope with the exponentially expanding body of publications in their field, while NLP researchers have been spurred to address new conceptual problems in theirs. Among the fundamental advances from the NLP perspective has been the realisation that tools which perform well on textual data from one source may fail to do so on another unless they are tailored to the new source in some way. This has led to significant interest in the idea of contrasting domains and the concomitant problem of domain adaptation, as well as the production of manualloy annotation domain-specific corpora. One definition of domain variation associates it with differences in the underlying probability distributions from which different sets of data are drawn (Xxxxx III and Xxxxx, 2006). The concept also mirrors the notion of variation across thematic subjects and the corpus-linguistic notions of register and genre (Biber, 1988). In addition to the differences in vocabulary that one would expect to observe, domains can vary in many linguistic variables that affect NLP systems. The scientific domain which has received the most attention (and is the focus of this paper) is the biomedical domain. Notable examples of corpus construction projects for the biomedical domain are PennBioIE (Xxxxxx et al., 2004) and XXXXX (Xxx et al., 2003). These corpora have been used to develop systems for a range of processing tasks, from entity recognition (Xxx et al., 2006) to parsing (Xxxx et al., 2005) to coreference resolution (Xxxxxx and Xxx, 2008). An implicit assumption in much previous work on biomedical NLP has been that particular subdomains of biomedical literature – typically molecular biology – can be used as a model of biomedical language in general. For example, XXXXX consist...
Related Papers. This work was presented at COLING 2012.
Related Papers. This work was presented at ACL 2012.
Related Papers. This work was presented at *SEM 2012. • X Xxxxxxxx, Xxxxxxxx and Xxxxxxxx, Xxxx. (2012). Modelling selectional preferences in a lexical hierarchy. In Proceedings of *SEM, Montreal, Canada. This paper describes Xxxxxxxx selectional preference models that incorporate knowledge from a lexical hierarchy such as WordNet. Inspired by previous work on modelling with WordNet, these approaches are based either on “cutting” the hierarchy at an appropriate level of generalisation or on a “walking” model that selects a path from the root to a leaf. In an evaluation comparing against human plausibility judgements, we show that the models presented here outperform previously proposed comparable WordNet-based models, are competitive with state-of-the-art selectional preference models and are particularly well- suited to estimating plausibility for items that were not seen in training. Recent research has investigated the potential of Bayesian probabilistic models such as Latent Dirichlet Allocation (LDA) for modelling selectional preferences (O S e´ aghdha, 2010; Xxxxxx et al., 2010; Xxxxxxxxx and Xxxxxx, 2011). These models are flexible and robust, yielding superior performance compared to previous approaches. In this paper we present a preliminary study of analogous models that make use of a lexical hierarchy (in our case the WordNet hierarchy). We describe two broad classes of probabilistic models over WordNet and how they can be implemented in a Bayesian framework. The two main potential advantages of incorporating WordNet information are: (a) improved predictions about rare and out-of-vocabulary arguments;
Related Papers. This work gave birth to two conference papers which provide more details about the implementation: • Quochi V., Xxxxxxxx F., Xxxxxx F. (2012) “A MWE Acquisition and Lexicon Builder Web Service”. Proceedings of the COLING 2012. Mumbai. India. • Frontini F., Xxxxxx X., Xxxxxx F. (2012) “Automatic Creation of quality Multi- word Lexica from noisy text data” Proceedings of the Sixth Workshop on Analytics for
Related Papers. The work using DTs has been presented at COLING 2010, on recognizing non-deverbal event nouns, and at LREC 2012, on recognizing a wide variety of noun classes.
Related Papers. This work was presented at CICLing 2011 and published in a proceedings volume. • Xxxxx, Xxxxx; Xxxxxxx, Xxxxxxx; Xxxxxx Xxxxxxxxx. (2011). Recognizing deverbal events in context. International Journal of Computational Linguistics and Applications – IJCLA Vol.2 (1-2).
