Abstract: According to an aspect, a candidate lexical kernel unit that includes a word token sequence having two or more words is received. Domain terms that contain the two or more words are retrieved from a terminology resource file of domain terms associated with a domain. The candidate lexical kernel unit and the retrieved domain terms are analyzed to determine whether the candidate lexical kernel unit satisfies specified criteria for use as a building block by a natural-language processing (NLP) tool for building larger lexical units in the domain. Each of the larger lexical units includes a greater number of words than the candidate lexical kernel unit. The candidate lexical kernel unit is identified as a lexical kernel unit based on determining that the candidate lexical kernel unit satisfies the specified criteria. The lexical kernel unit is output to a domain-specific lexical kernel unit file for input to the NLP tool.

Abstract: A predicate template collector allowing efficient and automatic recognition of predicate templates is adapted to include: a noun pair collector 94 and a noun pair polarity determiner 98 for collecting noun pairs co-occurring with predicate template pairs and determining polarity of relation between nouns, using conjunctions and seed templates; a template pair collector 100, collecting template pairs co-occurring with noun pairs and determining, based on the relation of noun pairs co-occurring with the predicate template pairs and the conjunctions between predicate templates pairs, whether the polarity of excitatory class of predicate template pair is the same or not; a template network builder 106 building a template network connecting predicate templates based on the predicate template pairs and match/mismatch of excitatory class thereof; and a template excitation value calculator 112 calculating excitation value to be assigned to each node, using the excitation value of seed templates and the relation betwe

Abstract: According to an aspect, a candidate lexical kernel unit that includes a word token sequence having two or more words is received. Domain terms that contain the two or more words are retrieved from a terminology resource file of domain terms associated with a domain. The candidate lexical kernel unit and the retrieved domain terms are analyzed to determine whether the candidate lexical kernel unit satisfies specified criteria for use as a building block by a natural-language processing (NLP) tool for building larger lexical units in the domain. Each of the larger lexical units includes a greater number of words than the candidate lexical kernel unit. The candidate lexical kernel unit is identified as a lexical kernel unit based on determining that the candidate lexical kernel unit satisfies the specified criteria. The lexical kernel unit is output to a domain-specific lexical kernel unit file for input to the NLP tool.

Abstract: One embodiment provides a system for recommending products. The system may include a search engine to retrieve, from a collection of product reviews, product review results using at least two input product names. The system may also include a template builder to build comparative sentence templates to define relationships between at least two product names, in which each comparative sentence template has a weight as a function of the defined relationship, and the search engine extracts one or more snippets matching at least one of the comparative sentence templates within each product review result. The system may further include a review ranking device to rank the product review search results based on the one or more extracted snippets, in which each snippet has a same weight as its matched comparative sentence template.

Abstract: A system is disclosing having a portable device, a computer, a network, and a display system. The portable device includes an input means for entering data and a wireless transceiver for transmitting the data and receiving messages. The computer includes an application and a display, the application includes a filter that checks the data against a list of words and indicates on the display if the data contains at least one of the words. The network is operatively connected to the portable device and the computer. The display system is connected to the computer and includes a display screen where the data is automatically displayed on the display screen for a predetermined period of time.

Abstract: Methods, systems, and apparatus are described for inducing a user of a speech recognition system to adjust their own behavior. For example, in one implementation, a speech recognition system that allows children to control electronic devices can improve the child's speech development, by encouraging the child to speak more clearly. To do so, the speech recognition system can generate a phonetic representation of a term spoken by the child, and can determine whether the phonetic representation matches a particular canonical pronunciation of the particular term that is deemed age-appropriate for the child. Upon determining that the particular canonical pronunciation that matches the phonetic representation of the term spoken by the child is not age-appropriate, the speech recognition system can select and implement a variety of remediation strategies for inducing the child to repeat the term using a pronunciation that is considered age-appropriate.

Abstract: A system and method are provided for adding user characterization information to a user profile by analyzing user's speech. User properties such as age, gender, accent, and English proficiency may be inferred by extracting and deriving features from user speech, without the user having to configure such information manually. A feature extraction module that receives audio signals as input extracts acoustic, phonetic, textual, linguistic, and semantic features. The module may be a system component independent of any particular vertical application or may be embedded in an application that accepts voice input and performs natural language understanding. A profile generation module receives the features extracted by the feature extraction module and uses classifiers to determine user property values based on the extracted and derived features and store these values in a user profile. The resulting profile variables may be globally available to other applications.

Abstract: A computer-readable recording medium stores a consolidation support program that causes a computer to execute a process that includes acquiring an electronic message transmitted and received among apparatuses; converting the electronic message into a first electronic message based on a first consolidation rule selected from a consolidation rule group to convert the electronic message by substituting with a specific value, a character string that matches a specific pattern and is included in the electronic message; converting the electronic message into a second electronic message based on a second consolidation rule having a priority rank that is higher than that of the first consolidation rule; determining whether the first and the second electronic messages match; and producing when the first and the second electronic messages do not match, a restrictive condition that the first consolidation rule is to be selected after the second consolidation rule, from the consolidation rule group.

Abstract: Distinct names of merchant entities in a transaction processing database are automatically corrected to standard names of entities by identifying non-standard features from the distinct names that do not uniquely identify the standard names of entities, and processing each distinct name with a selected regular expression tailored to remove the non-standard features and convent the names to a standard name format. Fuzzy matching is used to identify standard names of entities corresponding to the standard name formats.

Abstract: A method, computer program product, and computer system for identifying, by a computing device, content in a document, wherein the content includes a language expression. A context of the language expression is determined from a defined range of the content in the document. An action item associated with the language expression is generated based upon, at least in part, the context of the language expression.

Abstract: Machine logic that automatically detects natural language processing (NLP) system annotation errors and correspondingly updates NLP annotators to prevent future erroneous annotations by performing the following steps: (i) determining that a first annotation error has occurred in an annotation of a corpus by the natural language processing system; (ii) generating a candidate set of annotation correction actions, where each annotation correction action of the set is adapted to prevent an occurrence of an error similar to the first annotation error by the natural language processing system; (iii) selecting an annotation correction action from the candidate set of annotation correction actions, based, at least in part, on a set of annotation correction confidence characteristics; and (iv) automatically applying the selected annotation correction action to the natural language processing system.

Abstract: Implementations of the present disclosure are directed to a method, a system, and a computer program storage device for detecting a language in a text message. A plurality of different language detection tests are performed on a message associated with a user. Each language detection test determines a set of scores representing a likelihood that the message is in one of a plurality of different languages. One or more combinations of the score sets are provided as input to one or more distinct classifiers. Output from each of the classifiers includes a respective indication that the message is in one of the different languages. The language in the message may be identified as being the indicated language from one of the classifiers, based on a confidence score and/or an identified linguistic domain.

Abstract: Contextual adaptation of documents automatically replaces words for synonyms that appear within context or topic whey they are being used. A machine learned topic modeling, trained by a set of documents representative of a target user is executed to determine topics of an input document, and to determine words in the document to replace based on determining the relevance of the words to the topics in the documents. An output document is generated based on the input document with the replaced words.

Abstract: Creating and processing a natural language grammar set of data based on an input text string are disclosed. The method may include tagging the input text string, and examining, via a processor, the input text string for at least one first set of substitutions based on content of the input text string. The method may also include determining whether the input text string is a substring of a previously tagged input text string by comparing the input text string to a previously tagged input text string, such that the substring determination operation determines whether the input text string is wholly included in the previously tagged input text string.

Abstract: Systems and methods for improved management of sentiments over conventional approaches are disclosed. Supervised approach is used to augment the rule-based approach for classification. Initially, sentiment evaluation is performed by the system using a rule based approach and an interface is provided to the user to give feedback on the correctness of evaluated sentiment. This feedback is used by the sentiment evaluation system to update the set of rule-based and also apply the supervised approach to train the classifier for evaluating complex posts.

Abstract: A method for processing a natural language input to a computerized system. The method parses the input to identify a query portion and a communication portion of the input. The system then determines an answer to the query portion, including identifying communication parameters from the communication portion. Upon determining the answer, the system prepares an answer to the communication and transmits that answer. If the answer requires information from a remote source, the system creates a subsidiary query to obtain that information and then submits the subsidiary query to the remote source. A response to the query is used to compose the answer to the query from the answer to the subsidiary query. If the system concludes that the query portion does not require information from a remote source, analyzing and answering the query locally.

Abstract: Methods and devices for adding hyperlink to text are disclosed: generating hyperlink word list and characteristic word list in advance, determining co-occurrence frequency with each hyperlink word; to each text X which to be added the hyperlink, word segmentation processing them respectively, extracting the hyperlink word occurred in the hyperlink word list and the characteristic word occurred in the characteristic word list from results of word segmentation, determining weights of each extracted hyperlink word and extracted characteristic word, getting final weights of each extracted hypertext link word according to the co-occurrence frequency of each extracted characteristic word and each extracted hyperlink word and the weights; descendingly sorting each extracted hyperlink word according to the final weights, adding hyperlink to first k hyperlink words, and K is positive integer. Applying the solution, it can improve the relativity of the added hyperlink and the text, and it is easy to implement.

Abstract: A control system for selectively operating a sound system provided in a walk-around character costume. e.g., to play audio files in an order defined by a dialog tree that can be navigated by a performer wearing the costume. The control system includes a trigger mechanism provided in the costume operable by a performer wearing the costume to generate and transmit at least one trigger signal to navigate through a dialog tree. A controller receives the trigger signal and transmits a control signal causing the sound system to playback pre-recorded dialog. The system includes a trigger processing module run by the controller for processing the trigger signal and, in response to such processing, for navigating pre-recorded audio for a show to select the set of pre-recorded dialog. The trigger mechanism includes a first and a second pressure sensor in a thumb and index finger of a glove of the costume.

Abstract: Embodiments relate to managing and executing interpreted language code. An aspect includes a parser configured to parse controlled language code against both a language grammar and for building an execution model comprising parsed known concepts. Another aspect includes a pattern matching engine configured to match patterns in the controlled language code with a patterns for a new concept. Yet another aspect includes a concept engine configured to create a new concept from the matched pattern and controlled language code and for saving the new concept into an object model and into a language grammar so that current and future parsing errors will not occur when recognizing the new concept.

Abstract: Methods and arrangements for evaluating a regular expression. Text strings are received. A regular expression is also received, the regular expression comprising a pattern for specifying and recognizing at least one text string from among the received text strings. There is generated, with respect to the received text strings, a data structure containing grams with positional information. The data structure is employed to evaluate the regular expression via identifying a subset of the text strings comprising at least one match for the given regular expression. Other variants and embodiments are broadly contemplated herein.

Abstract: A method includes hierarchically identifying occurrences of some of the words in the set of sentences; creating a first index for each of some of the words based on the upper hierarchy of occurrences identified for each word; receiving input of a queried word; hierarchically identifying occurrences of the queried word in the set of sentences; creating a second index based on the upper hierarchy of occurrences identified for the queried word; comparing the first index and the second index to calculate an estimated value for the number of occurrences of a word in the neighborhood of the queried word; and calculating the actual value of the number of occurrences of a word in the neighborhood of the queried word based on an upper hierarchy and lower hierarchy of the occurrences on condition that the estimated value is equal to or greater than a predetermined number.

Abstract: Ontologies are used to comprehend regular expressions, by selecting, based on a context relating to a domain of a regular expression, an ontology and an assertion base, parsing the regular expression to identify at least one fragment of the regular expression, identifying one or more assertions in the assertion base corresponding to one of the identified fragments, identifying, for each identified assertion, an associated node in the ontology, and returning, based on the associated nodes, a concept in the ontology as representing the associated fragment of the regular expression.

Abstract: Techniques for annotating an entity in a document corpus using cross-document signals. A method includes determining which documents in a document corpus mention an entity of interest, clustering the documents that mention an entity of interest according to a temporal signal, a structural signal and/or a content signal, thereby forming at least one cluster of documents, and annotating at least one document in the at least one cluster of documents by marking each occurrence of the entity in the at least one document.

Abstract: Ontologies are used to comprehend regular expressions, by selecting, based on a context relating to a domain of a regular expression, an ontology and an assertion base, parsing the regular expression to identify at least one fragment of the regular expression, identifying one or more assertions in the assertion base corresponding to one of the identified fragments, identifying, for each identified assertion, an associated node in the ontology, and returning, based on the associated nodes, a concept in the ontology as representing the associated fragment of the regular expression.

Abstract: Techniques, including systems and methods, for providing translated content are described and suggested herein. In an embodiment, translations for instances of content are received and reviewers provide votes regarding the translations' quality. Votes received for translations are utilized in order to determine scores for the translations. Reviewers and translators may be scored as well. Scores for translations are used in various processes for providing appropriate content to users.

Abstract: A terminal device includes: a memory that stores a candidate group including a plurality of conversion word candidates for an input character; and a processor that, if a display rank of any candidate in the candidate group is changed, determines whether the changed display rank is included in a predetermined range of display ranks and, if the changed display rank is included in the predetermined range of display ranks, lets a communication unit transmit the candidate group including the changed display rank to another terminal device.

Abstract: A digital device and a method for parsing a query, in particular a natural language query, and retrieving results from possibly multiple data sources such as relational databases or the Semantic Web. The method includes a parsing procedure for generating a graph-based logical representation of the query using semantically structured resources, consisting of a tokenizer, a node generator, a relationship generator, and a focus identificator. The digital device realizes a modularized architecture, consisting of a parser enabling the processing of a query with possibly multiple vocabularies, a query performer retrieving data of knowledge sources independently from their database management system, and a result processor merging the results.

Abstract: A system for managing dictionaries, such as an application dictionary and a domain dictionary, and for adding entries to a data log is described herein. The system may, in response to a determination that an event occurs at an application, determine that the application uses a first concept name from the application dictionary to describe the event. An entry for the event may be added to a data log for the application. The entry may also include the first concept name from the application dictionary. A mapping of the first concept name from the application dictionary to a second concept name from the domain dictionary may be generated. In some aspects, the data log may be sent to a data log analysis system capable of accessing the domain dictionary.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for generating language models. In some implementations, data is accessed that indicates a set of classes corresponding to a concept. A first language model is generated in which a first class represents the concept. A second language model is generated in which second classes represent the concept. Output of the first language model and the second language model is obtained, and the outputs are evaluated. A class from the set of classes is selected based on evaluating the output of the first language model and the output of the second language model. In some implementations, the first class and the second class are selected from a parse tree or other data that indicates relationships among the classes in the set of classes.

Abstract: According to one embodiment of the invention, a method includes generating a person-name Information Gain (IG)-Tree and a relation IG-Tree from annotated data. The method also includes tagging and partial parsing of an input document. The names of the persons are extracted within the input document using the person-name IG-tree. Additionally, names of organizations are extracted within the input document. The method also includes extracting entity names that are not names of persons and organizations within the input document. Further, the relations between the identified entity names are extracted using the relation-IG-tree.

Abstract: A hybrid speech translation system whereby a wireless-enabled client computing device can, in an offline mode, translate input speech utterances from one language to another locally, and also, in an online mode when there is wireless network connectivity, have a remote computer perform the translation and transmit it back to the client computing device via the wireless network for audible outputting by client computing device. The user of the client computing device can transition between modes or the transition can be automatic based on user preferences or settings. The back-end speech translation server system can adapt the various recognition and translation models used by the client computing device in the offline mode based on analysis of user data over time, to thereby configure the client computing device with scaled-down, yet more efficient and faster, models than the back-end speech translation server system, while still be adapted for the user's domain.

Abstract: Data files are received from data sources that include textual content. The data files are categorized using a taxonomy of categories, where each category has sample textual content that defines a concept for the category. The categorizing includes comparing the textual content of the data file with the sample textual content for the category. A file score is calculated for each data file to compare the degree of similarity between the defined concept of the category and a determined concept for the data file. Each data file is associated with the category if the file score is equal to or greater than a pre-determined minimum score for the category. A portion of the data file and/or file score is be provided.

Abstract: A semantic tagging method may add context to a sentence in order to increase search efficiency. Regardless of an author's writing style, translating semantic concepts into tags may increase search efficiency. Automatic semantic tagging of documents may allow semantic search and reasoning. Text for semantic tagging may include an email, a website chat room, an internet forum, or a text message. Additional texts may include aggregating general consensus of an emailed topic across multiple emails, whether in the same email chain or separate emails. To increase search efficiency, the analysis of prior communications within the body of text may comprise analyzing structured contextual information to facilitate with homophora resolution. The structured contextual information may include at least one of a sender email address, one or more recipient email addresses, a subject field, a message date and time stamp, and an attachment title.

Abstract: In general, embodiments of the present invention provide systems, methods and computer readable media for generating an optimal classifying query set for categorizing and/or labeling textual data based on a query subsumption calculus to determine, given two queries, whether one of the queries subsumes another. In one aspect, a method includes generating a group of determining queries based on analyzing text within a document; receiving a group of classifying queries; and, for each determining query within the group of determining queries, determining whether at least one of the classifying queries is subsumed by the determining query; and updating the group of classifying queries in an instance in which the classifying query is subsumed by the determining query.

Abstract: Disclosed herein are methods and systems for recognizing speech. A method embodiment comprises comparing received speech with a precompiled grammar based on a database and if the received speech matches data in the precompiled grammar then returning a result based on the matched data. If the received speech does not match data in the precompiled grammar, then dynamically compiling a new grammar based only on new data added to the database after the compiling of the precompiled grammar. The database may comprise a directory of names.

Abstract: A request receiver (101) receives a request in which image ID information and a parameter for image processing are specified, from a terminal. An image processor (102) acquires an image based on the ID information that was specified in the received request, and by using the parameter that was specified in the received request to apply the image processing on the acquired image, outputs extracted information that was extracted from inside the image. A response transmitter (103) transmits a response in which the extracted information that was outputted by the image processor (102) is specified, to the terminal. When the extracted information that is associated with the ID information and the parameter that were specified in the received request is cached, the image processor (102), instead of acquiring the image and applying the image processing, obtains the cached extracted information, and outputs the obtained extracted information.

Abstract: A system and method for automated dictionary population is provided to facilitate the entry of textual material in dictionaries for enhancing word prediction. The automated dictionary population system is useful in association with a mobile device including at least one dictionary which includes entries. The device receives a communication which is parsed and textual data extracted. The text is compared to the entries of the dictionaries to identify new words. Statistical information for the parsed words, including word usage frequency, recency, or likelihood of use, is generated. Profanities may be processed by identifying profanities, modifying the profanities, and asking the user to provide feedback. Phrases are identified by phrase markers. Lastly, the new words are stored in a supplementary word list as single words or by linking the words of the identified phrases to preserve any phrase relationships. Likewise, the statistical information may be stored.

Abstract: A system, method, and computer-readable instructions for a distributed machine learning system are provided. A plurality of distributed learning environments are in communication over a network, wherein each environment has a computing device having a memory and a processor coupled to the memory, the processor adapted implement a learning environment via one or more agents in a rules-based system, wherein the agents learn to perform tasks in their respective learning environment; and a persistent storage in which knowledge comprising a plurality of rules developed by the agents for performing the tasks are stored, wherein the knowledge is tagged and shared with other agents throughout the plurality of distributed learning environments.

Abstract: In one embodiment, a method includes: receiving data representing a plurality of corpora, each of the plurality of corpora including a set of documents; receiving data representing terms that appear in the corpora; for each one of the terms, determining a plurality of inverse document frequency values each associated with a respective one of the plurality of corpora; receiving data representing a subset of the terms that also appear in a document; for each term in the subset, determining a term frequency for the term in the document; and for each term in the subset, determining, an augmented term frequency-inverse document frequency value based on: (i) the term frequency, and (ii) the plurality of inverse document frequency values that were determined for the term in the subset.

Abstract: Language dictation recognition systems and methods for using the same. In at least one exemplary system for analyzing verbal records, the system comprises a database capable of receiving a plurality of verbal records, the verbal record comprising at least one identifier and at least one verbal feature and a processor operably coupled to the database, where the processor has and executes a software program. The processor being operational to identify a subset of the plurality of verbal records from the database, extract at least one verbal feature from the identified records, analyze the at least one verbal feature of the subset of the plurality of verbal records, process the subset of the plurality of records using the analyzed feature according to at least one reasoning approach, generate a processed verbal record using the processed subset of the plurality of records, and deliver the processed verbal record to a recipient.

Abstract: A computer-implemented method manages data content, the data content including static data and dynamic data. The content management method includes analyzing the data content by a server to identify the static data and the dynamic data within the data content, wherein the dynamic data is included within a content block, and extracting the dynamic data in the content block from the data content, and replacing the content block with a special identifier. The content management method further includes hosting, by the server, the data content with the static content and the special identifier from a primary website in a primary domain, and the dynamic data in a secondary website in a secondary domain, wherein the special identifier identifies a location of the dynamic data.

Abstract: System and methods are disclosed for providing answers to search queries, and for searching using association data without requiring keyword matching. Datasets representing objects and their properties are created from unstructured data sources based on natural language analysis methods, and can be used to answer queries about objects or properties of objects. Implementations include general information search engines and embodiments for searching products, services, people, or other objects without knowing the names of such objects, or searching for information about known objects by using either keyword-based queries or natural language queries such as asking questions. System and methods are also provided for creating a structured or semi-structured representation of various unstructured data, in contrast to the conventional term-vector or term-document matrix representation.

Abstract: In some embodiments, a security device controller (SDC) is provided. In some embodiments, a security device controller includes receiving a configuration policy in a vendor neutral language; and automatically configuring a plurality of security devices on a heterogeneous network based on the configuration policy. For example, the plurality of security devices can include physical, virtual, or software defined network (SDN) based routers and/or firewalls, and the heterogeneous network can include security devices from a plurality of different vendors.

Abstract: For context-based audio filter selection, a type module determines a recipient type for a recipient process of an audio signal. The recipient type includes a human destination recipient type and a speech recognition recipient type. A filter module selects an audio filter in response to the recipient type.

Abstract: To appropriately execute a function based on a plurality of words, a function-execution instruction server of a function-execution instruction system includes: a function-execution instruction unit that issues an instruction of the execution of one or more tasks; a word input unit that inputs information containing a plurality of words that are arranged in order; and an executed-function determination unit that determines a task the execution of which is instructed on the basis of the order of words input.

Abstract: Methods, apparatuses, and computer program products are described herein that are configured to perform referring expression generation. In some example embodiments, a method is provided that comprises identifying an intended referent to be referred to in a textual output. The method of this embodiment may also include determining that a salient ancestor of the intended referent is lower in a part-of hierarchy than a lowest common ancestor. The method of this embodiment may also include causing the salient ancestor to be set as a current target referent and a new salient ancestor to be determined for the current target referent. In some example embodiments, the default descriptor of each current target referent is added to the referring noun phrase and the part-of hierarchy is traversed via salient ancestor links until the new salient ancestor of the current target referent is higher than or equal to the lowest common ancestor.

Abstract: A system is disclosing having a portable device, a computer, a network, and a display system. The portable device includes an input means for entering data and a wireless transceiver for transmitting the data and receiving messages. The computer includes an application and a display, the application includes a filter that checks the data against a list of words and indicates on the display if the data contains at least one of the words. The network is operatively connected to the portable device and the computer. The display system is connected to the computer and includes a display screen where the data is automatically displayed on the display screen for a predetermined period of time.

Abstract: The invention enables creation of grammar networks that can regulate, control, and define the content and scope of human-machine interaction in natural language voice user interfaces (NLVUI). The invention enables phrase-based modeling of generic structures of verbal interaction to be used for the purpose of automating part of the design of such grammar networks. Most particularly, the invention enables such grammar networks to be used in providing a voice-controlled user interface to human readable text data that is also machine-readable (such as a Web page, a word processing document, a PDF document, or a spreadsheet).

Abstract: An apparatus comprising: a processor; and a memory including computer program code, the memory and the computer program code configured to, with the processor, cause the apparatus to perform the following: detect, from a key user interface of an electronic device, key input of a particular key character from a first set of key characters; detect, from a user-scribe delineation user interface of the electronic device, user-scribed delineation input of a particular delineation character from a second set of delineation characters; wherein the first set of characters comprise grapheme characters and the second set of characters comprises embellishment marks, and wherein the apparatus is configured to associate together the particular key character and the particular delineation character for display.

Abstract: A log message is processed. The log message to be processed is received. One or more portions of the log message to be separately extracted are identified. A value is extracted from each identified portion. Extracting the value includes using an extraction rule. The extraction rule is associated with the identified portion.