Automatic Arabic Dialect Classification
The paper presents work on automatic Arabic dialect classification and proposes machine learning classification method where training dataset consists of two corpora. The first one is a small corpus of manually dialectannotated instances. The second one contains big amount of instances that were grabbed from the Web automatically using word-marks—most unique and frequent dialectal words used as dialect identifiers. In the paper we considered four dialects that are mostly used by Arabic people: Levantine, Egyptian, Saudi and Iraq. The most important benefit of that approach is the fact that it reduces time expenses on manual annotation of data from social media, because the accent is made on the corpus created automatically. Best results that we got were achieved with Naïve Bayes classifier trained using character-based bigrams, trigrams and word-marks vocabulary: precision of classification reaches 0.92 with F1 -measure equal to 0.91 on the test set of instances taken from manually annotated corpus.
This paper presents an algorithm that allows the user to issue a query pattern, collects multi-word expressions (MWEs) that match the pattern, and then ranks them in a uniform fashion. This is achieved by quantifying the strength of all possible relations between the tokens and their features in the MWEs. The algorithm collects the frequency of morphological categories of the given pattern on a unified scale in order to choose the stable categories and their values. For every part of speech, and for all of its categories, we calculate a normalized Kullback-Leibler divergence between the category’s distribution in the pattern and its distribution in the corpus overall. Categories with the largest divergence are considered to be the most significant. The particular values of the categories are sorted according to a frequency ratio. As a result, we obtain morphosyntactic profiles of a given pattern, which includes the most stable category of the pattern, and their values.
The paper describes the structure and possible applications of the theory of K-representations (knowledge representations) in bioinformatics and in the development of a Semantic Web of a new generation. It is an original theory of designing semantic-syntactic analyzers of natural language (NL) texts with the broad use of formal means for representing input, intermediary, and output data. The current version of the theory is set forth in a monograph by V. Fomichov (Springer, 2010). The first part of the theory is a formal model describing a system consisting of ten operations on conceptual structures. This model defines a new class of formal languages – the class of SK-languages. The broad possibilities of constructing semantic representations of complex discourses pertaining to biology are shown. A new formal approach to developing multilingual algorithms of semantic-syntactic analysis of NL-texts is outlined. This approach is realized by means of a program in the language PYTHON.
This paper is an overview of the current issues and tendencies in Computational linguistics. The overview is based on the materials of the conference on computational linguistics COLING’2012. The modern approaches to the traditional NLP domains such as pos-tagging, syntactic parsing, machine translation are discussed. The highlights of automated information extraction, such as fact extraction, opinion mining are also in focus. The main tendency of modern technologies in Computational linguistics is to accumulate the higher level of linguistic analysis (discourse analysis, cognitive modeling) in the models and to combine machine learning technologies with the algorithmic methods on the basis of deep expert linguistic knowledge.
The book contains the proceedigs of the 18th International Conference on Automatic Processing of Natural Langage (France, Montpellie, 27th June - 1st July 2011).
A framework for fast text analysis, which is developed as a part of the Texterra project, is described. Texterra provides a scalable solution for the fast text processing on the basis of novel methods that exploit knowledge extracted from the Web and text documents. For the developed tools, details of the project, use cases, and evaluation results are presented.
This paper concerns discourse-new mention detection in Russian. This might be helpful for different NLP applications such as coreference resolution, protagonist identification, summarization and different tasks of information extraction to detect the mention of an entity newly introduced into discourse. In our work, we are dealing with the Russian where there is no grammatical devices, like articles in English, for the overt marking a newly introduced referent. Our aim is to check the impact of various features on this task. The focus is on specific devices for introducing a new discourse prominent referent in Russian specified in theoretical studies. We conduct a pilot study of features impact and provide a series of experiments on detecting the first mention of a referent in a non-singleton coreference chain, drawing on linguistic insights about how a prominent entity introduced into discourse is affected by structural, morphological and lexical features.
This workshop is about major challenges in the overall process of MWE treatment, both from the theoretical and the computational viewpoint, focusing on original research related to the following topics:Manually and automatically constructed resources Representation of MWEs in dictionaries and ontologies MWEs in linguistic theories like HPSG, LFG and minimalism MWEs and user interaction Multilingual acquisition Multilingualism and MWE processing Models of first and second language acquisition of MWEs Crosslinguistic studies on MWEs The role of MWEs in the domain adaptation of parsers Integration of MWEs into NLP applications Evaluation of MWE treatment techniques Lexical, syntactic or semantic aspects of MWEs
he paper presents a framework for fast text analytics developed during the Texterra project. Texterra is a technology for multilingual text mining based on novel text processing methods that exploit knowledge extracted from user-generated content. It delivers a fast scalable solution for text mining without the expensive customization. Depending on use-cases Texterra could be utilized as a library, extendable framework or scalable cloudbased service. This paper describes details of the project, use-cases and results of evaluation for all developed tools. Texterra utilizes Wikipedia as a primary knowledge source to facilitate text mining in arbitrary documents (news, blogs, etc). We mine the graph of Wikipedia’s links to compute semantic relatedness between all concepts described in Wikipedia. As a result, we build a semantic graph with more than 5 million concepts. This graph is exploited to interpret meanings and relationships of terms in text documents. In spite of large size, Wikipedia doesn’t contain information about many domain-specific concepts. In order to increase applicability of the technology we developed several automatic knowledge extraction tools. These tools include systems for knowledge extraction from MediaWiki resources and Linked Data resources, as well as system for knowledge base extension with concepts described in arbitrary text documents using original information extraction techniques. In addition, utilization of information from Wikipedia allows easily extend Texterra for support of new Natural languages. The paper presents evaluation of Texterra applied for different text processing tasks (part-of-speech tagging, word sense disambiguation, keyword extraction and sentiment analysis) for English and Russian.