|
We have developed an unique Machine Translation (MT) tools
Morphology
English morphology contains flexions for all potentially inflected parts of speech such as nouns, adjectives, adverbs, verbs and pronouns. These are arranged into paradigms and mapped to respective grammatical characteristics: number, tense, person, etc.
There are 158 paradigms for verbs, 67 paradigms for nouns, 28 paradigms for adjectives and adverbs, 2 paradigms for pronouns (this & that) and paradigms for personal pronouns (I, me, my, mine, myself) will be added to the new version. Verb paradigms cover all regular and irregular verbs, and national spelling variants for US and UK English (travel -> traveled, travelled.) Noun paradigms cover popular & erudite forms for plural (cactus -> cactuses, cacti.) Paradigms for adjectives/adverbs cover simple, comparative and superlative forms and adverbs derived from adjectives, such as sharp -> sharper, sharpest, sharply. Paradigms for personal pronouns added to the new version cover subject, object, possessive, absolute possessive and reflexive forms.
Spanish morphology contains flexions for all potentially inflected parts of speech such as nouns, adjectives, verbs, pronouns, numerals and articles. All flexions are arranged into paradigms and mapped to respective grammatical characteristics such as number, gender, tense, person, etc. There are 53 paradigms (+ 22 paradigms for infixes) for verbs and 48 paradigms (+ 16 paradigms for infixes) for non-verbs. Verb paradigms cover all types of verb flexions, including paradigms for irregular verbs, regular and irregular participles (extraer -> extraído, extracto). Infixes are for alterations within word bases such as pensar -> pienso. Non-verb paradigms cover, among others, short forms for adjectives, pronouns and numerals such as bueno -> buen; alguno -> algún; ciento -> cien, adverbs derived from adjectives such as alto -> altamente and superlative degree for adjectives and adverbs like alto -> altísimo, altísima, altísimos, altísimas, altísimamente. Infixes are for alterations within word bases: imagen -> imágenes; fuerte -> fortísimo.
Polish morphology contains flexions for all potentially inflected parts of speech. All flexions are arranged into paradigms and are mapped to respective grammatical characteristics (number, gender, case, tense, person, etc.).
There are 403 paradigms for nouns and pronouns, 387 paradigms for verbs and 113 paradigms for adjectives. Noun paradigms cover all types of noun flexions, including optional forms (astrolog -> nom.pl.: astrolodzy, astrologowie; przecinek -> gen.sg.: przecineka, przecineku). Verb paradigms cover all types of verb flexions combining, wherever possible, perfect and imperfect forms in one paradigm (kupic-kupowac). Paradigms for adjectives cover, among others, adverbs derived from adjectives, comparative degree forms for adjectives and adverbs, including optional forms (nowy -> nowo, (po)nowie, nowszy, nowiej).
|
Russian, French, Italian, and/or Portuguese morphology could be developed and delivered within 2-6 months, depending on the language. Morphology for other languages could be developed upon request.
|
Syntactic analysis
A special metalanguage and syntactic rule system has been developed for analyzing English texts. The rule interpreter analyzes grammatical contexts (sets of grammatical charcteristics encoded in the dictionary), and tags and groups words in a sentence according to their syntactic roles, making all possible disambiguations and grammatical transformations, if necessary.
All rules are arranged and processed by priority. The rule system comprises such syntactic roles as noun group, verb group, kernel word in the group, adverb, attribute, determiner, agent, subject, subject complement, direct object, indirect object, predicate, predicate modifier, negation. Currently, the rule system covers about 40% of English test phrases and it could be developed further. Also, rule systems for syntactic analysis for other languages could be developed upon request. Estimated development time is about 6 months.
Text synthesis
A special metalanguage and rule system could be developed for reproducing text in Spanish (or in any other target language). The rule interpreter analyzes syntactic roles and grammatical characteristics encoded in the dictionary or assigned by the syntactic analyzer, and composes grammatically correct phrases in the target language, making all possible rearrangements and grammatical transformations, if necessary. Estimated development time is about 6 months.
Automatic transcriber for TTS
A special meta language and rule system has been developed for transcribing English words automatically. The rule interpreter analyzes grapheme composition of a word and produces its phonetic transcription, suitable for use in a TTS engine, with proper assignment of stress marks whenever possible. To increase accuracy, the transcriber uses a list of exceptions and a special algorithm that processes prefixes and suffixes in order to transcribe compound words properly. Rule systems for other languages could be developed upon request. Estimated development time is 1 to 6 months, depending on the language.
To obtain more information about licensing and
acquisition of our MT tools, contact us now.
|
 |
