One of the thorniest problems of large vocabulary continuous speech recognition systems is the large number of out-of-vocabulary (OOV) words. This is especially the case for the languages like Japanese, which has many inflections, compound words and loanwords. The OOV words vary with the application domains. It's not realistic to have a big general-purpose lexicon including any possible 00V wor...

The paper presents a corpus study which shows that the probability distribution of bi-phones in a lexicon of Korean loanwords is significantly different from that in a typical Korean lexicon or a lexicon consisting solely of native Korean and Sino-Korean words. This is demonstrated by comparing the perplexity of two types of bi-phone phonotactic models: a model trained on a set of Korean loanwo...

This paper deals with the robust expansion of Domain LexicoTaxonomy (DLT). DLT is a domain taxonomy enriched with domain lexica. DLT was proposed as an infrastructure for crossing domain barriers (Huang et al. 2004). The DLT proposal is based on the observation that domain lexica contain entries that are also part of a general lexicon. Hence, when entries of a general lexicon are marked with th...

In this demo paper, we present SARE, a modular and extendable semi-automatic system that 1) assists researchers in building gold-standard lexicons and evaluating their lexicon extraction algorithms; and 2) provides a general and extendable sentiment analysis environment to help researchers analyze the behavior and errors of a core sentiment analysis engine using a particular lexicon.

In this paper we propose an architecture for a lexicon management tool MANAGELEX. This tool aims at a general environment for reading, updating and combining lexicons in different formats. The starting point is the already existing lexicon models MULTILEX and GENELEX. Each functionality (reading, updating and combining) is based on a corresponding model, which can be configured and maintained c...

We present a novel algorithm for the acquisition of multilingual lexical taxonomies (including hyponymy/hypernymy, meronymy and taxonomic cousinhood), from monolingual corpora with minimal supervision in the form of seed exemplars using discriminative learning across the major WordNet semantic relationships. This capability is also extended robustly and effectively to a second language (Hindi) ...

This paper deals with several lexicographic problems arising from the choice of hypernyms in the Dictionary of Medieval Scientific French (Créalscience database). The generic names used by medieval scholars indicate categorizations which differ from modern classifications. We examine some cases selected from the medieval domain of natural sciences where we can notice a conceptual discrepancy be...

We describe a method for creating a nonEnglish subjectivity lexicon based on an English lexicon, an online translation service and a general purpose thesaurus: Wordnet. We use a PageRank-like algorithm to bootstrap from the translation of the English lexicon and rank the words in the thesaurus by polarity using the network of lexical relations in Wordnet. We apply our method to the Dutch langua...

The paper reports on the development and experimental implementation of a combined methodology of knowledge transfer from an existing lexical resource with a view to semiautomatically inducing a new lexicon, in a cross-language environment. The source resource, a multi-layered Italian lexicon whose theoretical approach to the content and representation of semantic information follows the genera...