Abbyy Infopoisk LLC

Abbyy Software Limited

Abbyy Production LLC

Abbyy Development Inc.

Findo, Inc.

Konstantin Anisimovich

Konstantin Zuev

Vladimir Pavlovich Selegey

Diar Tuganbaev

Vladimir Rybkin

Tatiana Danielyan

Dmitry Deryagin

Vadim Tereshchenko

David Yan

Alexander Shamis

Evgeny Indenbom

Sergey Kolotienko

Evgenii Mikhailovich Indenbom

Dmitry Vnuchkov

Ivan Ivanovich Ivashnev

Andrey Anatolievich Isaev

Anatoly Starostin

Stepan Evgenyevich Matskevich

Tatiana Parfentieva

Aleksey Zhuravlev

Ruslan Garashchuk

Andrey Valerievich Myakutin

Anton Krivenko

Valery Igorevich Novitskiy

Evgeny Shavlyugin

Azat Aydarovich Davletshin

Andrey Lubenets

Sergey Platonov

Aleksei Zhuravlev

Embodiments of the present disclosure describe a system and method for optical character recognition. In one embodiment, a system receives an image depicting text. The system extracts features from the image using a feature extractor. The system applies a first decoder to the features to generate a first intermediary output. The system applies a second decoder to the features to generate a second intermediary output, wherein the feature extractor is common to the first decoder and the second decoder. The system determines a first quality metric value for the first intermediary output and a second quality metric value for the second intermediary output based on a language model. Responsive to determining that the first quality metric value is greater than the second quality metric value, the system selects the first intermediary output to represent the text.

Optical character recognition using a combination of neural network models

Systems and methods for automatic generation of templates for information extraction rules to extract information objects from natural language texts. An example method of automatic generation of templates for information extraction comprises receiving, by a computer device, a first text fragment, comprising a first identifier of a first text token, wherein the first token comprises one or more words in a natural language and wherein the first token references a first information object from a first category of information objects; displaying, using a template-generating graphical user interface, a plurality of linguistic characteristics of the first token; receiving, via the template-generating graphical user interface, a first input identifying template attributes from the plurality of the linguistic characteristics of the first information object; generating a first template based at least in part on the identified template attributes; creating a first production rule for the first template; applying the first production rule to portions of a first natural language text matching the first template; displaying, using the template-generating graphical user interface, a second information object identified in the first natural language text using the first production rule.

System and method of automatic template generation

A classification engine generates, using a weighted graph, a plurality of sets of confused graphemes based on recognition data for a plurality of document images; receives an input grapheme image associated with a document image comprising a plurality of grapheme images; determines a set of recognition options for the input grapheme image, where the set of recognition options comprises a set of target characters that are similar to the input grapheme image; identifies a neural network trained to recognize a first set of confused graphemes, where the first set of confused graphemes comprises at least a portion of the set of recognition options for the input grapheme image; and determines a grapheme class for the input grapheme image using the identified neural network.

Differential classification using multiple neural networks

Systems and methods for using autoencoders for training natural language classifiers. An example method comprises: producing, by a computer system, a plurality of feature vectors, wherein each feature vector represents a natural language text of a text corpus, wherein the text corpus comprises a first plurality of annotated natural language texts and a second plurality of un-annotated natural language texts; training, using the plurality of feature vectors, an autoencoder represented by an artificial neural network; producing, by the autoencoder, an output of the hidden layer, by processing a training data set comprising the first plurality of annotated natural language texts; and training, using the training data set, a text classifier that accepts an input vector comprising the output of the hidden layer and yields a degree of association, with a certain text category, of a natural language text utilized to produce the output of the hidden layer.

Using autoencoders for training natural language text classifiers

An illustrative method according to a set of instructions stored on a memory of a computing device includes receiving, by a processor of the computing device, a search input. The method further includes searching, by the processor, a plurality of electronic storage locations for electronic files related to the search input. The method further includes determining, by the processor, the presence of an electronic file on a first electronic storage location of the plurality of electronic storage locations that is related to the search input. The method further includes displaying, by the processor, on a graphical user interface (GUI), representative information of the electronic file. The representative information includes descriptive information relating to the electronic file.

Method and system for cross-platform searching of multiple information sources and devices

Systems and methods for evaluating text classifier parameters based on semantic features. An example method comprises: performing a semantico-syntactic analysis of a natural language text of a corpus of natural language texts to produce a semantic structure representing a set of semantic classes; identifying a natural language text feature to be extracted using a set of values of a plurality of feature extraction parameters; partitioning the corpus of natural language texts into a training data set comprising a first plurality of natural language texts and a validation data set comprising a second plurality of natural language texts; determining, in view of the category of the training data set, the set of values of the feature extraction parameters; validating the set of values of the feature extraction parameters using the validation data set.

Evaluating text classifier parameters based on semantic features

Systems and methods for multi-stage recognition of named entities based on morphological and semantic features of natural language texts. An example method comprises: performing a lexico-morphological analysis of a natural language text comprising a plurality of tokens, each token comprising at least one natural language word; determining, based on the lexico-morphological analysis, one or more lexical meanings and grammatical meanings associated with each token of the plurality of tokens; for each token the plurality of tokens, evaluating one or more classifier functions using the lexical and grammatical meanings associated with the tokens, wherein a value of each classifier function is indicative of a degree of association of the token with a category of named entities; performing a syntactico-semantic analysis of at least part of the natural language text to produce a plurality of semantic structures representing the part of the natural language text; and interpreting the semantic structures using a set of production rules to determine, for one or more tokens comprised by the part of the natural language text, a degree of association of the token with a category of named entities.

Multi-stage recognition of named entities in natural language text based on morphological and semantic features

An example method for natural language text classification based on semantic features comprises: performing semantico-syntactic analysis of a natural language text to produce a semantic structure representing a set of semantic classes; associating a first semantic class of the set of semantic classes with a first value reflecting a specified semantic class attribute; identifying a second semantic class associated with the first semantic class by a pre-defined semantic relationship; associating the second semantic class with a second value reflecting the specified semantic class attribute, wherein the second value is determined by applying a pre-defined transformation to the first value; evaluating a feature of the natural language text based on the first value and the second value; and determining, by a classifier model using the evaluated feature of the natural language text, a degree of association of the natural language text with a category of a pre-defined set of categories.

Natural language text classification based on semantic features

A method and computer system for translating sentences between languages from an intermediate language-independent semantic representation is provided. On the basis of comprehensive understanding about languages and semantics, exhaustive linguistic descriptions are used to analyze sentences, to build syntactic structures and language independent semantic structures and representations, and to synthesize one or more sentences in a natural or artificial language. A computer system is also provided to analyze and synthesize various linguistic structures and to perform translation of a wide spectrum of various sentence types. As result, a generalized data structure, such as a semantic structure, is generated from a sentence of an input language and can be transformed into a natural sentence expressing its meaning correctly in an output language. The method and computer system can be applied to in automated abstracting, machine translation, natural language processing, control systems, Internet information retrieval, etc.

Method and system for translating sentence between languages based on semantic structure of the sentence

Various embodiments are provided for enabling tagging of image files based on tagging or commenting on images on a social networking site. The tagging or commenting on an image on the social networking site is detected by the system. The social network tag or comment is analyzed to determine a textual tag to be assigned to image files corresponding to the social network images that have been tagged or commented on. In some implementations semantic analysis of text component of the social network tags or comments is performed. In some implementations the textual tags are then propagated to other image files associated with the user.

Tagging of images based on social network tags or comments

The preferred embodiments provide an automated machine translation from one language to another. The source language may contain expressions or words that are not readily handled by the translation system. Such problematic words or word combinations may, for example, include the words not found in the dictionary of the translation system, as well as text fragments corresponding to structures with low ratings. To improve translation quality, such potentially erroneous words or questionable word combinations are identified by the translation system and displayed to a user by distinctive display styles in the display of a document in the source language and in its translation to a target language. A user is provided with a capability to correct erroneous or questionable words so as to improve the quality of translation.

Indicating and correcting errors in machine translation systems

Disclosed are methods, systems, and computer-readable mediums for automatic training of a syntactic and semantic parser using a genetic algorithm. An initial population is created, where the initial population comprises a vector of parameters for elements of syntactic and semantic descriptions of a source sentence. A natural language compiler (NLC) system is used to translate the sentence from the source language into a target language based on the syntactic and semantic descriptions of the source sentence. A vector of quality ratings is generated where each quality rating in the vector of quality ratings is of a corresponding parameter in the vector of parameters. Quality ratings are evaluated according to specific criterion, which comprise parameters such as a BLEU score and a number of emergency sentences. A number of parameters in the vector of parameters are replaced with adjusted parameters.

Automatic training of a syntactic and semantic parser using a genetic algorithm

A system for natural language processing is provided. A first natural language processing program may be constructed using language-independent semantic descriptions, and language-dependent morphological descriptions, lexical descriptions, and syntactic descriptions of one or more target languages. The natural language processing program may include any of machine translation, fact extraction, semantic indexing, semantic search, sentiment analysis, document classification, summarization, big data analysis, or another program. Additional sets of natural language processing programs may be constructed.

Exhaustive automatic processing of textual information

The present disclosure provides methods and systems for performing syntactic analysis of a text. In some implementations the method includes performing rough syntactic analysis of the text, generating a graph of generalized constituents of the text and filtering arcs of the graph of generalized constituents with a combination classifier which includes a tree classifier and one or more linear classifiers. The combination classifier is trained using parallel analysis of an untagged two-language text corpus.

Growing community of inventors

Moscow, Russia

Konstantin Anisimovich