Abstract: A facility for generating a graphic image is described. The facility receives from a user a body of text whose creation is not subject to any rules or prototypes. The facility discerns from the body of text a textual organization. The facility then generates a graphic image conveying the discerned textual organization.

Abstract: In an embodiment a device to convert conversations from a meeting to text and annotate the text is disclosed. In an embodiment a device is disclosed, comprising: a microphone; a camera; a processor; and a storage comprising a set of instructions; wherein the set of instructions causes a processor to: receive from the microphone, an audio recording containing speech of a participant of a meeting; receive from the camera, a video of the participant; identify the participant; convert the speech of the participant to a digital text; develop a skeletal map of the participant; recognize a gesture of the participant from the skeletal maps; detect and identify a target of the gesture; based on the target and the gesture determine an annotation for the digital text corresponding to a point of time of the gesture.

Abstract: A method and system for annotating and linking electronic documents is described herein. Separate annotation layers or connectors are used to store each annotation or connection associated with a document in order to improve the efficiency and robustness of collaborative annotation. When a user creates an annotation for a document, a new annotation layer is generated, containing information describing the annotation. The annotation layer is separate from the annotated document. The annotation layer may be transmitted to another user without transmission of the document, thereby reducing network traffic and avoiding metadata contamination within the document itself. This facilitates real-time collaborative annotation of electronic documents by multiple users. This also facilitates robust connections between documents or other data sources, which contain information regarding both the source and target documents.

Abstract: System for converting a spreadsheet in a first format to a second format comprising: a server including a storage means adapted to store the spreadsheet in the second format and store an associated template; a data set associated with the spreadsheet in the second format, such that the data set is adapted to be associated with the template to be displayed on a client device; and wherein when data is associated with the template, changes made via the client device to the associated data is recorded in realtime in both the spreadsheet in the second format and template. A process for generating an output for a webpage, comprising: selecting a spreadsheet and a template stored by a server, the spreadsheet comprising a data set; associating the data set with an element in the template, such that when the data is associated with the template, changes made to the associated data is recorded in realtime in both the spreadsheet and template; and generating the output comprising the data associated with the template.

Abstract: Mechanisms are provided for implementing a logical reasoning and justification engine that operates to receive a logical parse data structure of natural language content. The logical parse data structure comprises nodes and edges linking nodes and identifies latent logical terms within the natural language content indicative of logical relationships between elements of the natural language content. The engine further operates to receive a selection of a node in the logical parse data structure to thereby form a selected node, and execute at least one logical justification module on the selected node to identify zero or more justifying nodes that provide a contribution to a knowledge state of the selected node. The engine further operates to generate a logical justification output based on the identified zero or more justifying nodes, and output the logical justification output.

Abstract: A method and system for determining a location of origin and a time period in which a document was written is disclosed. A text is received and a set of linguistic characteristics for the text are identified. A set of possible locations and time periods for the text are determined based on the set of linguistic characteristics. A set of reference documents are used to determine a proximity rating for the text based upon a determination of how close the text is to the reference documents. The potential locations and time periods are ranked and returned for presentation.

Abstract: Each of a plurality of content matchers is executed upon a body of text in a document, identifying at least one match in the text and additionally assigning a match strength for each match. Where a conflict between a first match (have a first match strength associated therewith) and a second match (having a second match strength associated therewith) is noted, it is determined whether either of the first or the second match strength is greater than the other. If so, that match of the first and second matches corresponding to the lesser of the first and second match strengths is discarded. If the first or second match strengths are equal, then respective matcher ranks of the first matcher and the second matcher are compared such that the match of the first and second matches corresponding to the lesser of the first and second matcher ranks is discarded.

Abstract: Embodiments described herein of the natural language determiner (NLD) illustrate the ability to determine, at runtime, the conversion of natural language text into a computer-usable form.

Abstract: A method for entity recognition employs document-level entity tags which correspond to mentions appearing in the document, without specifying their locations. A named entity recognition model is trained on features extracted from text samples tagged with document-level entity tags. A text document to be labeled is received, the text document being tagged with at least one document-level entity tag. A document-specific gazetteer is generated, based on the at least one document-level entity tag. The gazetteer includes a set of entries, one entry for each of a set of entity names. For a text sequence of the document, features for tokens of the text sequence are extracted. The features include document-specific features for tokens matching at least a part of the entity name of one of the gazetteer entries. Entity labels are predicted for the tokens in the text sequence with the named entity recognition model, based on the extracted features.

Abstract: At least two processing device-implemented company name recognition components, operating upon a body of text in a document, identify at least one company name occurrence in the body of text based at least in part on a company identifier list. The company name recognition techniques implemented by each of the at least two company name recognition components are different from each other. The at least one company name occurrence is used to update the company identifier list. The updated company identifier list is then used by the at least two company name recognition components to identify at least one additional name occurrence in the same body of text. This process of repeatedly identifying occurrences of company names in the body of text and updating the company identifier list is performed until such time that no further company name occurrences are identified in the body of text.

Abstract: An information processing apparatus to select a token from a document to describe a field of interest includes an obtaining unit, a determining unit, a clustering unit, and a selecting unit. The obtaining unit obtains a list of tokens output from extractors that received the document as an input. Each output token has an extractor score assigned to by an extractor. The determining unit determines, as a word frequency value, a frequency of each word in the list of tokens, determines a token score for each token in the list of tokens, and determines a distance between each token in the list of tokens. The clustering unit clusters each token in the list of tokens into a plurality of groups. The selecting unit selects a token with a group of the plurality of groups to describe the field of interest in the document.

Abstract: Textual Geographical Location relates a placename, which is a set of terms, from one to any maximum as defined in an individual language, to a unique point or area (many points) as found on a map or other coordinate system, such as the map of the United States as used in global positioning system (GPS).

Abstract: An emotion estimator of a dialogue control device estimates a dialogue partner's emotion based on the dialogue partner's facial expression or voice prosody. A dialogue controller controls the dialogue with the dialogue partner using a topic selected based on whether the emotion estimated by the emotion estimator is good or bad. Specifically, the dialogue controller controls the dialogue with the dialogue partner based on a topic map created by a topic mapper and including a topic that the dialogue partner likes.

Abstract: A device may obtain a test script document. The device may process the test script document to perform term extraction using one or more term extraction techniques to identify a set of terms of the test script document. The one or more term extraction techniques may include a skip n-gram term extraction technique. One or more terms, of the set of terms, may be located within an n-gram of the test script document. The device may process the test script document to perform hierarchy formation for results of performing term extraction. A relationship between a set of terms, of the set of terms, may be identified using hierarchy formation. The device may generate a functional diagram of the test script document based on the results of performing term extraction and results of performing hierarchy formation. The device may provide information identifying the functional diagram.

Abstract: Embodiments are used to extract terms from any text set that are used on other text, such as in a repository, that then can be used in a variety of applications, from providing search results, to analyzing data sets, to building a variety of text generation tools, such as messaging and emails.

Abstract: A device may obtain a document. The device may identify a skip value for the document. The skip value may relate to a quantity of words or a quantity of characters that are to be skipped in an n-gram. The device may determine one or more skip n-grams using the skip value for the document. A skip n-gram, of the one or more skip n-grams, may include a sequence of one or more words or one or more characters with a set of occurrences in the document. The sequence of one or more words or one or more characters may include a skip value quantity of words or characters within the sequence. The device may extract one or more terms from the document based on the one or more skip n-grams. The device may provide information identifying the one or more terms.

Abstract: At least one processing device, operating upon a body of text in a document, identifies occurrences of at least one content type in the body of text. The at least one processing device thereafter generates a user interface that includes portions of text from the body of text that are representative of at least some of the occurrences of the at least one content type in the document. For each content type, the occurrences corresponding to that content type can be grouped together to provide grouped content type occurrences that are subsequently collated in the user interface. Those portions of text corresponding to the grouped content type occurrences may be arranged in alphanumeric order. The user interface may comprise at least a portion of the body of text as well as indica indicating instances of the occurrences within the portion of the body of text.

Abstract: Systems, methods, and computer-readable media for providing semantically-relevant discovery of solutions are described herein. In some examples, a computing device can receive an input, such as a query. The computing device can process each word of the input sequentially to determine a semantic representation of the input. Techniques and technologies described herein determine a response to the input, such as an answer, based on the semantic representation of the input matching a semantic representation of the response. An output including one or more relevant responses to the request can then be provided to the requestor. Example techniques described herein can apply machine learning to train a model with click-through data to provide semantically-relevant discovery of solutions. Example techniques described herein can apply recurrent neural networks (RNN) and/or long short term memory (LSTM) cells in the machine learning model.

Abstract: An approach is provided to discover new portmanteau, such as when ingesting documents into a question answering (QA) system. The approach works by analyzing a words included in electronic documents and identifies words as being possible portmanteaus. To analyze a portmanteau found in a document, the approach identifies morphemes that are included in the identified portmanteau and candidate words that correspond to each of the identified morphemes. A meaning for the new portmanteau is then derived from the meanings of the candidate word meanings.

Abstract: A method includes receiving a text. The method also includes identifying a set of linguistic characteristics contained in the text. The method also includes determining a plurality of time periods in which the text was potentially written based on the set of linguistic characteristics. The method also includes retrieving a set of reference documents for each time period. The method also includes producing a set of proximity scores by performing a set of proximity checks using the set of linguistic characteristics, the set of reference documents, and the text, where the proximity checks analyze how often and how close linguistic characteristics are to one another. The method also includes ranking the plurality of time periods based on the set of proximity scores and returning a set of one or more ranked time periods of the plurality of time periods.

Abstract: A method includes receiving a text and identifying a set of linguistic characteristics contained in the text, where linguistic characteristics include grammatical, syntactic, and idiomatic features of the text. The method also includes determining a plurality of locations of origin in which the text was potentially written based on the set of linguistic characteristics. The method also includes retrieving a set of reference documents for each location of origin in the plurality of locations of origin and producing a set of proximity scores by performing a set of proximity checks using the set of linguistic characteristics, the set of reference documents, and the text, wherein the proximity checks analyze how often and how close linguistic characteristics are to one another. The method also includes ranking the plurality of locations of origin based on the set of proximity scores and returning a set of one or more ranked locations of origin.

Abstract: Systems and methods for training and using a natural language understanding system are provided. More specifically, the systems and methods train a natural language understanding system utilizing a distributed network of feature extractors on features servers. Further, the systems and methods for using the natural language understanding system utilize a distributed network of features extractor on features servers. Accordingly, the systems and methods provide for a more accurate natural langue understanding system, a more reliable natural langue understanding system, and a more efficient natural langue understanding system. Further, the systems and methods provide for natural language understanding systems with better development (including update ability), productivity, and scalability.

Abstract: An approach is provided to discover new portmanteau, such as when ingesting documents into a question answering (QA) system. The approach works by analyzing a words included in electronic documents and identifies words as being possible portmanteaus. To analyze a portmanteau found in a document, the approach identifies morphemes that are included in the identified portmanteau and candidate words that correspond to each of the identified morphemes. A meaning for the new portmanteau is then derived from the meanings of the candidate word meanings.

Abstract: The personal translator implementations described herein provide a speech translation device that pairs with a computing device to translate in-person conversations. The speech translation device can be wearable. In one implementation the personal translator comprises a speech translation device with at least one microphone that captures input signals representing nearby speech of a first user/wearer of the device and at least one other nearby person in a conversation in two languages; a wireless communication unit that sends the captured input signals representing speech to a nearby computing device, and receives for each language in the conversation, language translations from the computing device; and at least one loudspeaker that outputs the language translations to the first user/wearer and at least one other nearby person. The language translations in text form can be displayed on a display at the same time the language translations are output to the loudspeaker(s).

Abstract: Techniques are provided for automatically creating online accounts based on digital images, such as digital images of business cards. In one technique, multiple data items that have been extracted from a digital image of a business card are identified. A particular data item is contact information of a user associated with (or identified by) the business card. A verification code is sent, based on the particular data item, to a computing device of the user. The verification code is received from the computing device of the user. In response to receiving the verification code an account is created for the user and the account is modified to include a least some of the multiple data items.

Abstract: Techniques include outputting to a developer an offer to opt-in to a translation feature that enables human translators to translate their web page to a target language. In response to receiving a first request to opt-in to the translation feature, the server: generates and stores a web page copy, obtains from the human translators translations of at least a portion of the web page from its source language to the target language, modifies the web page copy based on the obtained translations to obtain a translated web page that is a translated version of the web page, detects a second request for the web page from a computing device associated with the target language, and in response to detecting the second request outputs, to the computing device, the translated web page with additional content relevant to the computing device or a user associated with the computing device.

Abstract: A mobile electronic device has a touch sensitive screen and an accelerometer. A translator is to translate a word or phrase that is in a first human language and that is entered via a first virtual keyboard displayed on the touch sensitive screen, into a second human language. A translator is to cause the touch sensitive screen to display the translated word or phrase and a second virtual keyboard having characters in the second human language, in response to the accelerometer detecting a change in the physical orientation of the device or movement of the device. Other embodiments are also described and claimed.

Abstract: A statistics-based machine translation method is disclosed. The method generates probabilities of translation from a sentence to be translated to candidate translated texts based on features of the candidate translated texts that affect the probabilities of translation and a pre-generated translation probability prediction model. The features that affect probabilities of translation include at least degrees of semantic similarity between the sentence to be translated and the candidate translated texts. A preset number of candidate translated texts with highly ranked probabilities of translation are selected to serve as translated texts of the sentence to be translated. The method is able to go deep into a semantic level of a natural language when a machine translation model is constructed to avoid a semantic deviation of a translated text from an original text, thereby achieving the effect of improving the quality of translation.

Abstract: Implementations herein relate to methods and devices for generating candidate translations and for quantizing text as well as words. A method may include generating, by a computing device, pending candidate translations of text to be translated based on predetermined translation rules. The computing device may generate translation probabilities from the text to be translated to the pending candidate translations based on features having impacts on translation probabilities of the pending candidate translations and a predetermined translation probability prediction model. The computing device may then select a predetermined number of pending candidate translations that have the translation probabilities higher than other pending candidate translations in the pending candidate translations to be the candidate translations of the text to be translated.

Abstract: The present invention relates to searching documents based on a range index. A range index may comprise range-searchable elements having explicit data types, such as integers, real numbers, geographic locations, dates, times, etc. The data type determines what operators can be used in expressions. Each element of the range index corresponds to an occurrence of an item in the document collection that satisfies the range expression. In addition, range indexes can be aggregated into a set of aggregate indexes to facilitate evaluation. Aggregation may depend on the field's type or the distribution of values. For example, date fields might have a range indexed by day, by month at one level, and then by year at another level. Thus, these new file structures and evaluation techniques result in a new inverted list structure that allows efficient query evaluation using expressions that operate on typed data.

Abstract: Aspects of the present innovations relate to systems and/or methods involving multimedia modules, objects or animations. According to an illustrative implementation, one method may include accepting at least one input keyword relating to a subject for the animation and performing processing associated with templates. Further, templates may generates different types of output, and each template may include components for display time, screen location, and animation parameters. Other aspects of the innovations may involve providing search results, retrieving data from a plurality of web sites or data collections, assembling information into multimedia modules or animations, and/or providing module or animation for playback.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method, and/or computer program product embodiments for a crowd sourced indexing and/or searching of content. An embodiment operates by receiving one or more requests for content from one or more media devices, each request comprising content identifier information that identifies the content, determining whether crowd sourced content index information has been generated for the content, transmitting a response to the one or more media devices of the one or more media devices, in response to the one or more requests, the response comprising content location information and a content indexing request, and receiving content index information for the content identified by the content identifier information from the one or more media devices.

Abstract: Approaches for routing data to storage are provided. An approach includes determining implicit metadata from explicit metadata received with a request from a user to store a file. The approach also includes determining a storage resource based on the explicit metadata, the implicit metadata, and a registry of storage resources. The approach additionally includes routing data of the file to the determined storage resource.

Abstract: Provided herein are mixed-media modules with enhanced features that can be used as search results. Systems and methods are disclosed for performing processing involved with search, such as processing search information to return search results. In one exemplary implementation, there is provided a method for processing information to return search results including mixed-media media presentation(s) selectable by a user. Moreover, such method may involve user interaction to manipulate the presentation, display various media and/or effect other functionality. Further implementations may involve generation of interactive, visually rich mixed-media content of high information density providing improved user experience and/or improved value to various participants.

Abstract: In an approach for creating a file, a computer receives a selection on a selection of content within application software on a first computing device to transfer to a clipboard. The computer appends metadata of the received selection with metadata indicating application software and file association information that corresponds to the selected content. The computer receives an indication to paste the selected content from the clipboard to a second computing device. The computer determines whether compatible application software on the second computing device is capable of accepting the selected content included in the received indication to paste. The computer responsive to determining that compatible application software on the second computing device is not capable of accepting the selected content included in the received indication to paste, creates a file that includes the selected content included in the received indication to paste and appended metadata.

Abstract: A content retrieval method including: extracting a plurality of fingerprints including a plurality of video fingerprints and audio fingerprints from contents stored in a content database; determining representative video fingerprints of the video frames and representative audio fingerprints of the audio sequences; determining a data rate indicating a storage limitation and a coverage indicating a number of searching results to be returned; storing selected representative video fingerprints and representative audio fingerprints based on the storage limitation indicated by the data rate in a fingerprint database; receiving a query containing at least one of video data and audio data submitted by a user; extracting at least one query fingerprint representing the query; determining a number of fingerprints most matching the query fingerprint based on the coverage to generate search results indicating matching contents represented by the number of most matching fingerprints; and returning the search results to the

Abstract: A method for a color search is disclosed. The method may be implemented by a user device terminal. The method may comprise sending a color search request and corresponding color encoding information to a network device, and receiving search result information that matches the color encoding information and that is returned by the network device based on the color search request.

Abstract: An apparatus includes a user interface module that displays a search user interface element to a user. The search element is associated with a multimedia object. An object type module determines type data for the multimedia object indicating an object type of the multimedia object. A query wrapper construction module creates a query wrapper based on the type data. A network communication module transmits the query wrapper to a search system and receives a result set. The result set includes identifying information of a first application state of a first application and a first access mechanism for the first application state. A result presentation module presents the result set to the user. An access module, in response to actuation of a first user interface element, opens the first application to the first application state according to the first access mechanism and provides the multimedia object to the first application state.

Abstract: Techniques are disclosed herein for controlling object placement in object storage. A placement component of a storage application receives a request to store a first object in an object store having multiple nodes. The object store determines a placement of the first object to one of the nodes based on an object namespace including a numerical namespace and a lexicographical namespace. Each node is assigned a corresponding subspace of the object namespace for storing objects. The first object includes a numerical namespace value and a lexicographical namespace value. A second object (a replica of the first object) is generated. The first object is stored to a first node based on the lexicographical namespace value. The second object is stored to a second node based on the numerical namespace value.

Abstract: A method, a computer system, and/or a computer program product are disclosed. One computer-implemented method for building a hash table includes dividing a hash table into plural blocks; and dividing each block into plural sub-blocks. A certain sub-block uses a first pattern of association between a key and a location for storing the key. Another sub-block which belongs to the same block having the certain sub-block uses a second pattern which is different from the first pattern. The method may further include building a hash table by using memory blocks in a Field Programmable Gate Array.

Abstract: A method of finding a match in a match database with a target media includes receiving a target media item, wherein the target media item comprises one or more pictures or videos, extracting multiple visual descriptors from the target media item, formulating descriptor vectors from the multiple visual descriptors, reducing the dimension of the descriptor vectors to generate projected descriptors, creating cluster keys from the projected descriptors, and generating a list of a number of matches by comparing the cluster keys of the target media item with keys of sequences stored in the matching database.

Abstract: Techniques are described for automated management of natural language queries of enterprise data. In one example, a method includes performing natural language processing and semantic processing on a natural language query to identify data sets relevant to the natural language query. The method further includes grouping the data sets into one or more query domains based at least in part on one or more relationships among the data sets. The method further includes prioritizing the query domain sets. The method further includes loading one or more of the query domain sets in an order based on the prioritizing of the query domain sets.

Abstract: Methods and apparatus for targeted message delivery are disclosed. A set of document delivery transactions are initiated with respect to a set of intended document recipients. The initiating includes sending to each of the set of intended document recipients a respective one of a set of pointers for requesting delivery of one or more documents designated with respect to a respective document delivery transaction. Each of the set of pointers is identified with a respective one of the set of intended document recipients. In response to actuations of respective ones of the set of pointers, the one or more documents designated with respect to the respective document delivery transaction are delivered from a centralized document repository. Analytic measures are recorded with respect to the delivering the one or more documents. The analytic measures include a measure of actuations of each of the pointers. A target for a subsequent communication is identified.

Abstract: A solution is provided to generate video recommendations in a video sharing environment. A video recommendation system selects a video as a target video and extracts target keywords from the title of the identified target video or the title of a non-video trending news item. The system receives multiple candidate videos. For each candidate video, the system extracts keywords from the title of the candidate video and compares the extracted keywords with the target words. Based on the comparison, the system generates a similarity score for the candidate video. The system ranks the candidate videos based their associated similarity scores and selects a candidate video having the highest similarity score as the video recommendation for the target video.

Abstract: A method for providing group-based attention management of electronic messages may include receiving an electronic message, the electronic message defining a plurality of recipients. One or more candidate tuples are then identified from the plurality of recipients. The one or more candidate tuples are compared with one or more stored tuples to identify one or more relevant tuples from the candidate tuples. Based on the one or more relevant tuples, an attention score for one or more of the plurality of recipients is determined. Based on the attention score for one or more of the plurality of recipients, the presentation of the electronic message to one or more of the plurality of recipients is modified.

Abstract: The technology described herein provides an efficient mechanism for quickly analyzing huge amounts of media content to find media content (hereafter “content” or “media content”) that is relevant to a user. The technology analyzes features of a curator to classify curators by interest and/or find curators with similar content recommendations. The curator data can be used to make curator recommendations to users based on the user's interests. The technology described herein collects curator data from multiple content sites and analyzes the data to identify curators that recommend similar content on different content sites.

Abstract: Techniques for generating an expert database and verifying an expert using member data are described. A search request can be received from a device of the user. The search request can include a specific skill associated with the expert. Additionally, profile data can be accessed from a database in the online social network. Additionally, an expert recommendation process can determine an expert from the members of the online social network based on the search request and the profile data of the members. Moreover, social graph data can be accessed from a second database in the online social network. Furthermore, a connection path process can determine the connection path between the user and the expert based on the social graph data. Subsequently, the determined expert and the determined optimal path between the user and the expert can be presented on a display of the device.

Abstract: A social network can use at least one computer processor for visually displaying performance of content on the social network. The computer processor can receive a selection of a time frame and an audience on a user interface provided to a user. The selected audience can include a plurality of users of the social network specified by category. The computer processor can retrieve, from the social network, topic data relating the content to user engagement of the content during the selected time frame by the selected audience. The content is categorized by topic. The computer processor can rank the topics, of the retrieved topic data, as a function of the user engagement. The computer processor can create a visual representation of the highest-ranked topics including indicia indicating relative levels of the user engagement for corresponding topics. The computer processor can display the visual representation on the user interface.

Abstract: A computer-implemented method for customising a travel itinerary for an account holder is provided. The method comprises receiving current transaction data when settling a current travel transaction initiated by the account holder with a current merchant; retrieving a plurality of historical transaction data relating to a plurality of historical transactions settled for the account holder prior to the current travel transaction; identifying one or more of the plurality of historical transactions that relate to the current transaction data to recognize preferences of the account holder; and customising a travel itinerary based on the preferences of the account holder, the travel itinerary including at least one merchant with whom the account holder can initiate a future transaction.

Abstract: Query requests for RDF triples are obtained, wherein the query request(s) contain(s) at least one triple pattern; for each triple pattern, the corresponding elementary pattern is determined, and each triple pattern is converted to a weighted elementary pattern. The occurrence frequency of each elementary pattern is computed based on the weighted elementary patterns; at least one elementary pattern is chosen at least according to the occurrence frequency; and the RDF triples corresponding to the chosen at least elementary pattern are prefetched into the buffer. The corresponding apparatus is also provided. With the above method and apparatus, the frequently accessed RDF triples can be determined and prefetched into the buffer, which improves the query efficiency.