Abstract: An approach to classify different defect records by mapping plain language phrases to a taxonomy. The approach includes a method that includes receiving, by at least one computing device, a defect record associated with a defect. The method further includes receiving, by the least one computing device, a plain language phrase or word. The method further includes mapping, by the least one computing device, the plain language phrase or word to a taxonomy. The method further includes classifying, by the least one computing device, how the defect was at least one of detected and resolved using the taxonomy.

Abstract: A method is disclosed for evaluating a chat message sent between users of an online environment. The method may include associating each word in the chat message with metadata. The metadata identifies a word type and usage for each word in the chat message. This method may also include identifying one or more safety rules associated with the metadata. Each safety rule identifies an ordered sequence of one or more sets of words. This method may also include applying the safety rule to the chat message to determine whether a sequence of words in the chat message present in the ordered sequence of sets of words. Upon determining a word, from each set of words in the ordered sequence of sets of words, matches a respective one of the words in the chat message, the chat message is blocked from being sent to a message recipient.

Abstract: Technology is disclosed for snippet pre-translation and dynamic selection of translation systems. Pre-translation uses snippet attributes such as characteristics of a snippet author, snippet topics, snippet context, expected snippet viewers, etc., to predict how many translation requests for the snippet are likely to be received. An appropriate translator can be dynamically selected to produce a translation of a snippet either as a result of the snippet being selected for pre-translation or from another trigger, such as a user requesting a translation of the snippet. Different translators can generate high quality translations after a period of time or other translators can generate lower quality translations earlier. Dynamic selection of translators involves dynamically selecting machine or human translation, e.g., based on a quality of translation that is desired.

Abstract: Mechanisms are provided for translating messages output by processes associated with the data processing system. Language based user activities occurring via a data processing system are tracked and automatically analyzed to determine a relative priority ordering of one or more languages used by a user performing the user activities. A sequence of preferred language translation substitutions is generated for outputting messages based on results of the analysis and is automatically applied to a received message from a process associated with the data processing system to generate a translated message using one of the preferred language translation substitutions in the sequence of preferred language translation substitutions. The translated message is output via an output device of the data processing system.

Abstract: A computer implemented method for detecting linguistic uniqueness includes receiving, with an examination evaluation server, a list of word n-grams and corresponding frequency values indicating the frequency at which each word n-gram appears in a language corpus, receiving an examination response with one or more sentences, tokenizing the sentences into a plurality of words or phrases, combining contiguous words or phrases into n-word segments, and identifying, for each n-word segment that matches a matching word n-gram from the list, the frequency value corresponding to the matching word n-gram.

Abstract: An expression image matching method includes obtaining an identity of a session in which a user chats, the session being a dialog set in which the user chats with at least one other user. The method determines, according to a white list and the identity of the session, whether to provide an expression image to the user, and the white list comprising an identity of a session in which a frequency of using expression images meets a predefined condition. The method provides an expression image matching with chat content input by the user when determining to provide the expression image to the user.

Abstract: A message processing device (10) includes an estimator (11), an extractor (12), and an indicator (13). The estimator (11) estimates words understandable to a destination user (1). The extractor (12) extracts, from a message (3) created by a transmission originator user (2), a portion that does not match the words estimated by the estimator (11). The indicator (13) indicates, to the transmission originator user (2), the message (3) with the extracted portion by the extractor (12) being in an emphasized manner.

Abstract: The invention relates to a system (SYS) for automatically extracting a location of an abnormality with respect to an anatomical structure from a report, the system comprising a tokenizer (U10) for tokenizing the report or a part of it, thereby producing a plurality of tokens, and an analyzer (U20) for identifying a semantic structure comprising identified tokens of the plurality of tokens, describing the location of the abnormality with respect to the anatomical structure. Optionally, the system further comprises a mapper (U30) for annotating a diagram representing the anatomical structure, based on the identified semantic structure describing the location of the abnormality with respect to the anatomical structure. Using the system, the location of the abnormality with respect to the anatomical structure can be extracted from each report of the plurality of reports produced over the period of time. The extracted locations may be used by a clinician for diagnostic purposes.

Abstract: The information replying method includes: receiving to-be-replied information, where the to-be-replied information includes text content and contact information; searching a database for corresponding dialog style information according to the text content and the contact information; performing preprocessing on the text content, where the preprocessing includes word segmentation processing and stop word removal processing; and searching, according to data that has undergone the preprocessing, the database corresponding to the dialog style information, to determine reply information.

Abstract: A technique relates comparing content. A first set of content in a first document and a second set of content in a second document are normalized. The first set of content in the first document and the second set of content in the second document are tokenized. The first set of content having been tokenized and the second set of content having been tokenized are compared in order to find differences in the second set of content with respect to the first set of content. The differences are compiled in a changes map. The differences in the changes map are analyzed to determine types of the differences in the first set of content and the second set of content, and predefined differences are to be excluded. A report of the differences is generated, and the report delineates the types of the differences while excluding the predefined differences.

Abstract: An apparatus state estimation method includes: storing first apparatus data that indicates a state of an apparatus; storing a work record, the work record being a record indicating a work that has been carried out for maintenance of the apparatus; estimating a category that is a malfunction of the apparatus or a content of a measure by using the work record; and learning a relationship between the estimated category and the first apparatus data.

Abstract: Systems and methods for aspect-based sentiment analysis using machine learning methods. An example method comprises: receiving, by a computer system, a custom dictionary comprising a list of lexemes referencing at least one of: a target entity or an aspect associated with the target entity; performing, using the custom dictionary, a syntactico-semantic analysis of at least part of a natural language text to produce a plurality of syntactico-semantic structures representing the part of the natural language text; evaluating, using one or more text characteristics produced by the syntactico-semantic analysis, a classifier function to determine polarities associated with one or more aspect terms; and generating a report comprising the aspect terms and polarities of aspects referenced by the aspect terms.

Abstract: The present invention relates to statistical machine translation, and provides a machine translation device and a machine translation method that acquire a creation probability for a target language from a single corpus while extracting respective conversion probabilities by extracting syntax conversion knowledge and word translation knowledge from a parallel corpus, model a weighted translation model by allowing each of the conversion knowledge and each of the probabilities to learn using a translation model learning device, and generate a target sentence through decoding processes of a syntax converter and a word translator by applying the translation model to a source sentence input in real time, thereby resolving disadvantages of the existing phrase-based SMT and syntax-based SMT and combining advantages thereof.

Abstract: Devices and methods for determining the content of a first segment of text in a first language, using a second segment of text in a second language. The second segment of text is a translation of the first segment of text.

Abstract: Aspects of the present invention provide a technique to validate the transfer of intents or entities between existing natural language model domains (hereafter “domain” or “NLU”) using click logs, a knowledge graph, or both. At least two different types of transfers are possible. Intents from a first domain may be transferred to a second domain. Alternatively or additionally, entities from the second domain may be transferred to an existing intent in the first domain. Either way, additional intent/entity pairs can be generated and validated. Before the new intent/entity pair is added to a domain, aspects of the present invention validate that the intent or entity is transferable between domains. Validation techniques that are consistent with aspects of the invention can use a knowledge graph, search query click logs, or both to validate a transfer of intents or entities from one domain to another.

Abstract: Methods for classifying a point of view and diegesis are provided. A method can include providing a processor in operable communication with a computer-readable medium, receiving a narrative text, extracting a set of features from the narrative text, transmitting the features into a feature vector, transmitting a plurality of feature vectors to a support vector machine, predicting a point of view and diegesis for the narrative text associate with a particular feature vector, and annotating the narrative text.

Abstract: Methods, systems, and computer program products for translating structured languages to natural language using domain-specific ontology are provided herein. A computer-implemented method includes determining similarities among multiple natural language query interpretations derived from an input query, determining differences among the multiple natural language query interpretations, and generating natural language descriptions of each of the multiple natural language query interpretations based on analysis of the determined similarities, the determined differences, and the input query. The method also includes producing, for each of the natural language query interpretations, a natural language string that represents one or more unambiguous interpretations of the input query, wherein the producing comprises consolidating the generated natural language descriptions. Further, the method includes outputting each of the produced natural language strings to a user.

Abstract: Described herein is an approach for automatically determining the semantic relatedness of documents to semantic concepts. A first text mining analysis extracts a set of reference concepts from reference documents. A second text mining analysis extracts a set of test concepts from test documents that include a mixture of new concepts and reference concepts. An extended co-occurrence matrix is computed that indicates a frequency of co-occurrence (RCCF) of each new and each reference concept in the test documents with all other new and reference concepts. The extended co-occurrence matrix is used for computing a new concept relatedness score (NCRS) for the new concepts. A document similarity score (DSS) is computed for each of the test documents by aggregating, inter alia, the NCRS of each new concept with the RCCF of each reference concept. The DSS represents the semantic relatedness of the test document to the totality of the reference concepts.

Abstract: A technique for translating text strings includes receiving a source language text string from an application, determining that a translated text string that includes a translation in a target language of the source language text string is not available for use by the application, transmitting the source language text string to a translation service for translation, receiving the translated text string from the translation service, and causing the translated text string to be available for use by the application.

Abstract: According to one aspect, there is provided an apparatus for improving a bilingual corpus including a plurality of sentence pairs of a first language and a second language and word alignment information of each of the sentence pairs, the apparatus comprises: an extracting unit for extracting a split candidate from word alignment information of a given sentence pair; a calculating unit for calculating split confidence of said split candidate; a comparing unit for comparing said split confidence and a pre-set threshold; and a splitting unit for splitting said given sentence pair at said split candidate in a case that said split confidence is larger than said pre-set threshold.