The Semantic Web, a Web with the meaning, is often associated with specific XMLbased standards for semantics, such as RDF and OWL [http://www.w3.org/RDF/, http://www.w3.org/TR/owl-features/]. If HTML and the Web made all the online documents look like one huge book, RDF, schema, and inference languages will make all the data in the world look like one huge database. One of the key promises of the Semantic Web is that it will provide the necessary infrastructure for enabling services and applications on the Web to automatically aggregate and integrate information into a sum which is greater than the individual parts. So the Semantic Web should enable users to locate, select, employ, compose, and monitor Web-based services automatically. To make use of a Web service a software agent needs a computerinterpretable description of the service, and the means by which it is accessed. An important goal for Semantic Web markup languages is to establish a framework within which these descriptions are made and shared. Web sites should be able to employ a standard ontology, consisting of a set of basic classes and properties, for declaring and describing services, while the cross-lingual ontology structuring mechanisms of OWL provide an appropriate, Web-compatible representation language framework within which to do this.
Web-compatible representation language framework today usially is based on lexical ontologies. Wordnets are cross-lingual lexical ontologies, including information on hypernyms, synonyms, polysemous terms, relations between terms, and sometimes multilingual equivalents. Wordnets are valuable resources as sources of ontological distinctions. WordNets provide a conceptual framework for multilingual mappings in ontologies. Linking concepts across many cross-lingual lexicons belonging to the WordNet-family started by using the Interlingual Index (ILI) [2]. Unfortunately, no version of the ILI can be considered a standard and often the various lexicons exploit different version of WordNet as ILI.
At the 3rd GWA Conference in Korea there was launched the idea to start building a WordNet grid around a Common Base Concepts expressed in terms of WordNet synsets and SUMO definitions (http://www.globalwordnet.org/gwa/gwa_grid.htm). This first version of the Grid was planned to be build around the set of 4689 Common Base Concepts. Since then only three languages with essentially various number of synsets and different WordNet versions were placed in the Grid mappings (English -4689 synsets with WN 2.0 mapping, Spanish -15556 synsets with WN1.6 mapping and Catalan -12942 synsets with WN1.6 mapping). But there is yet no official format for the Global WordNet Grid. So far there are just only 3 files in the specified format.
This paper reports about the current results of the English-Russian WordNet development [2,3,4]. It describes usage of Russian and English-Russian lexical language resources and software to process English-Russian WordNet and English-Russian WordNet Grid (4600 synsets with WN 3.0 mapping) and design of a XML/RDF/OWL-markup of the English-Russian WordNet resources. Relevant aspects of the DTD/XML/RDF/OWL formats and related technologies are surveyed.
On December 2003 our research group got license from OUP to explore and exploit for research purposes such language resources: -Oxford Russian Dictionary; -New Oxford Dictionary of English, 2nd Edition; -New Oxford Thesaurus of English.
For many linguistic tasks of WordNet development we use language processor Russicon that includes such main blocks:
System allows receiving morphological information of the word and to build normal form of the word, shows paradigm for the word, constructs new words lexicon, constructs frequency lexicon (Fig. 1).
The theoretical foundation of the morphological analyzer and normalyzer program is a language-independent model of morphological analysis [6][7][8]. Morphological analyzer and normalyzer allows a) defining the following grammatical characteristic s of a word: part of speech, case, gender, number, tense, person, degree of comparison, voice, aspect, mood, form, type, transitiveness, reflexive, animation, b) modifying a given word to its normal grammatical form/s -lemma/s.
WordNet editor TenDrow was developed to help join production of Russian WordNet from above mentioned linguistic resources. It allows to join sysnsets from Thesaurus, explanatory and other dictionaries; proceed relations between synsets and words of synsets. WordNet editor is not only viewer but also a real tool for constructing and editing multiligual/monolingual WordNet. It is a database management system in which users (linguist or knowledge engineer) can create, edit and look at the English and Russian (Fig. 2).
Usually there were several standard variants (Fig. 3, a,b,c,d) of the English-Russian WordNet and English-Russian WordNet Grid translation equivalents.
The simplest is the a variant. Approximately 24000 English-Russian synsets could be translated in such way. The hardest is d variant because such kind of translation destroys normal mapping and forms additional sub mappings. More then 15000 English synsets have no right word to word translation to Russian. Synset: {sell} -be sold at a certain price or in a certain way Example: These books sell like hot cakes.
Russian translation: {prodavat'sya;rasprodavat'sya;sbyvat'sya} Hyponomy problems: sometime no translations of English synset member exists in Russian or there were some loops in relations : Synset: {сutter, cutlery, cutting tool} -a cutting implement; a tool for cutting.
Hypernym: ENG20-03040079-n Synset: {cutting implement} -a tool used for cutting or slicing.
Russian translation: { kolyusche-rejuschie orudiya} Hypernym: ENG20-04279652-n Synset: {edge tool} -any cutting tool with a sharp cutting edge (as a chisel or knife or plane or gouge).
Hypernym: ENG20-03039706-n
The porting of the English-Russian WordNet was done into XML using the DTD for the XML structure from http://www.globalwordnet.org/gwa/gwa_grid.htm and the The English-Russian DTD and XML format for the English-Russian WordNet and English-Russian WordNet Grid is shown on Fig. 4,5. The WordNet Task Force [9] developed a new approach in WordNet RDF conversion. The W3C WordNet project is still in the process of being completed, at the level of schema and data (http://www.w3.org/2001/sw/BestPractices/WNET/wn-conversion.html). It was used for porting of the English-Russian WordNet and English-Russian WordNet Grid into RDF and OWL.
But still there are open issues how to support different versions of WordNet in XML/RDF/OWL and how to define the relationship between them and how to integrate WordNet with sources in other languages. ?aWordSense wn20schema:word ?aWord . ?aWord wn20schema:lexicalForm "bank"@en-US } Proposed semantic framework [8] for grid improvement is based on such main counterparts (Fig. 6): RDF/OWL store; tools for information extraction; tools for Ontology Engineering Modeling Process; knowledge mining, SPAROL/SQL search and analysis tools.