Abstract: In some implementations, data indicating a touch received on a proximity-sensitive display is received while the proximity-sensitive display is presenting one or more items. In one aspect, the techniques describe may involve a process for disambiguating touch selections of hypothesized items, such as text or graphical objects that have been generated based on input data, on a proximity-sensitive display. This process may allow a user to more easily select hypothesized items that the user may wish to correct, by determining whether a touch received through the proximity-sensitive display represents a selection of each hypothesized item based at least on a level of confidence that the hypothesized item accurately represents the input data.

Abstract: The present invention is directed to a method, medium and system for use in a computing environment for managing input information. The system for facilitating input management in a computerized environment may include an input manager for processing events received from an input provider. The system may additionally include a staging area including a stack of events created by the input manager. The staging area allows access by external components. The external components may include a set of filters capable of accessing and manipulating the stack of input events and a set of monitors capable of monitoring manipulation of the stack of events.

Abstract: A forms processing method involves accessing an individual scanned form, analyzing the individual form, based upon the analysis, selecting a proper forms-scanning template, loading multiple field-specific dictionaries linked to individual fields in the selected forms-scanning template, analyzing, text content, on a field by field basis using the field-specific dictionary linked to each respective field, for misspellings or improper entries, attempting to identify whether a closest fit between the particular content and contents of the field-specific dictionary linked to that field exists and, if the closest fit exists, automatically replacing the particular content with an entry from the field-specific dictionary linked to that field corresponding to the closest fit.

Abstract: In an approach for creating a hierarchal organization of spoken conference call topics, a processor determines an identity of a first speaker from a group of participants on a conference call, wherein the group of participants are known. A processor determines a first conference call topic, wherein the first conference call topic is determined using a portion of the first speaker's speech. A processor adds a second speaker's identity to a queue of speakers while recording the first speaker's voice. A processor determines a second conference call topic. A processor links the first conference call topic and the second conference call topic into a hierarchal organization of spoken conference call topics.

Abstract: A method of searching a business listing with voice commands includes receiving, over the Internet, from a user terminal, a query spoken by a user, which includes a speech utterance representing a category of merchandize, a speech utterance representing a merchandize item, and a speech utterance representing a geographic location.

Abstract: Techniques for generating a query statement to query a dataset may be provided. For example, the query statement can be generated from natural language input, such as a natural language utterance. To do so, the input can be analyzed to detect a sentence, identify words in the sentence, and tag the words with the corresponding word types (e.g., nouns, verbs, adjectives, etc.). Expressions using the tags can be generated. Data about the expressions can be inputted to a classifier. Based on a detected pattern associated with the expressions, the classifier can predict a structure of the query statement, such as what expressions correspond to what clauses of the query statement. Based on this prediction, words associated with the expressions can be added to the clauses to generate the query statement and accordingly query the dataset.

Abstract: A method, performed on a server, of translating between languages includes receiving first audio data for a first language from a mobile device, translating the first audio data to second audio data for a second language, receiving an indication that the mobile device has moved between two locations, and sending the second audio data to the mobile device in response to the indication.

Abstract: An information processing device processes target information described using identifiers formed from a symbol, a product operator that joins together factors containing one or more of the identifiers and thereby forms a string of ordered factors, and a sum operator that joins together terms containing one or more of the factors and thereby forms an expression that is a combination of the terms. Also, information processing device includes a dividing unit that divides the target information into a plurality of terms at a predetermined position; and a transmission unit that associates positional information with at least any of the factors contained in the terms resulting from the division and causes each of the terms resulting from the division to be held in any of a plurality of server devices, where the positional information contains an occurrence sequence of the terms in the target information and an occurrence sequence of the factors in the terms.

Abstract: System and method to characterize a security entity in a computing environment is disclosed. Communication between a user computer and at least one destination computer by a security appliance is monitored by a security appliance. Selective information from the communication is extracted by the security appliance, selective information indicative of the security entity. A plurality of words from the communication between the identified security entity and at least one destination computer are selectively extracted. A word cloud is generated based on the selectively extracted plurality of words. The word cloud is evaluated to characterize the identified security entity.

Abstract: Methods and apparatus for speech recognition on user dictated words to generate a dictation and using a discriminative statistical model derived from a deterministic formatting grammar module and user formatted documents to extract features and estimate scores from the formatting graph. The processed dictation can be output as formatted text based on a formatting selection to provide an integrated stochastic and deterministic formatting of the dictation.

Abstract: A method, apparatus, and system are directed towards seeding a user's contacts for their online social network. The invention is arranged to automatically recommend to the user a set of seed contacts that the user may employ to invite to join their social network. The set of seed contacts may be harvested from the user's existing portal activities, as well as other sources. In one embodiment, the invention analyzes portal activity, such as email exchanges with the user, and the like, to determine a frequency of contact with the user. Other sources may include but not be limited to emails, names within an address book of the user, names within an address book of another person, a buddy list, an instant messaging list, an activity, a mailing list, an online discussion group, a membership in a category, chat group, and the like.

Abstract: Because an information providing system according to the present invention generates recognition object words for selecting one headline from a plurality of headlines in such a way that the recognition object words do not overlap one another, the information providing system can select a headline uniquely through the user's utterance, and present information (content) corresponding to the headline selected thereby, and therefore the convenience to the user is improved.

Abstract: One or more computing devices may receive audio data from a first client device. The one or more computing devices may also receive a designation of a second client device from the first client device. The one or more computing devices may transcribe the audio data to text, and may further identify profile information associated with a user of the first client device in the transcribed text. The profile information may be stored to a profile associated with the user of the first client device. The one or more computing devices may also transmit at least one of the audio data or the transcribed text to the second client device.

Abstract: An apparatus for normalizing input data of an acoustic model includes a window extractor configured to extract windows of frame data to be input to an acoustic model from frame data of a speech to be recognized, and a normalizer configured to normalize the frame data to be input to the acoustic model in units of the extracted windows.

Abstract: Methods and systems for a transportation vehicle are provided. For example, one method includes receiving a user input by a processor executable, digital assistant at a device on a transportation vehicle; determining by the digital assistant a relationship between words of the user input; using a trained dataset by the digital assistant to ascertain an intended action based on the user input; and providing a response by the digital assistant for the user input, where the response includes performing the intended action using another device on the transportation vehicle.

Abstract: User interactions with a computing system are sensed and recorded. The recording represents a process for controlling a computer system. Voice input are received and the computer system actions that are taken based upon the voice inputs and a task recording.

Abstract: Technologies are disclosed herein for statistical machine translation. In particular, the disclosed technologies include extensions to conventional machine translation pipelines: the use of multiple domain-specific and non-domain-specific dynamic language translation models and language models; cluster-based language models; and large-scale discriminative training. Incremental update technologies are also disclosed for use in updating a machine translation system in four areas: word alignment; translation modeling; language modeling; and parameter estimation. A mechanism is also disclosed for training and utilizing a runtime machine translation quality classifier for estimating the quality of machine translations without the benefit of reference translations.

Abstract: A voice recognition device includes an extractor configured to extract at least one of a first utterance element indicating an execution command and a second utterance element indicating a subject, from a user's voice utterance, a domain determiner configured to determine a current domain to provide response information regarding the voice utterance based on the first and the second utterance elements, and a controller configured to determine a candidate conversation frame to provide the response information regarding the voice utterance on at least one of the current domain and a previous domain based on a conversation state of the current domain, wherein the previous domain is determined from a previous voice utterance of the user.

Abstract: An electronic device and a method for controlling the same are provided. The electronic device includes a storage configured to store domain information that is categorized for dialog subjects, a speaker configured to output a system response based on a user utterance sound, and a processor configured to detect a domain, among the domain information, based on the user utterance sound, determine one among the detected domain and a previous domain as a domain to be used to process the user utterance sound, based on a confidence between the user utterance sound and the detected domain, and process the user utterance sound to generate the system response, based on the determined domain.

Abstract: Multilingual information guidance system and device are provided. To elaborate, the system may include: at least one information output device configured to display guidance information to a user; and an information provision server connected to the information output device and configured to provide guidance information prepared in one or more languages to the information output device according to a request of the information output device. The information output device may output guidance information prepared in the native language corresponding to nationality information of a user, which is read from an RFID tag in which the nationality information of the user is stored or transmitted from the information provision server.

Abstract: Method of selecting training text for language model, and method of training language model using the training text, and computer and computer program for executing the methods. The present invention provides for selecting training text for a language model that includes: generating a template for selecting training text from a corpus in a first domain according to generation techniques of: (i) replacing one or more words in a word string selected from the corpus in the first domain with a special symbol representing any word or word string, and adopting the word string after replacement as a template for selecting the training text; and/or (ii) adopting the word string selected from the corpus in the first domain as the template for selecting the training text; and selecting text covered by the template as the training text from a corpus in a second domain different from the first domain.

Abstract: A communication device (100) comprising a controller (210), a radio frequency interface (230) and a memory (240), wherein the controller is configured to identify at least one receiving communication device (100B, 100C); present content (510); receive one user input (U) indicating a share; and share the content (510) with the at least one receiving communication device (100B, 100C) through the sharing network over the radio frequency interface (230); wherein the communication device (100) is characterized in that the controller (210) is further configured to identify the receiving communication device (100B, 100C) by the controller (210) being configured to detect a presence of the receiving communication device (100B, 100C) being in close proximity and then initiate a connection with the at least one receiving communication device (100B, 100C) and wherein the controller is configured to share content upon initiating the connection thereby establishing a sharing network (400) automatically.

Abstract: A method, system and computer program product for presenting tags of a tag cloud in a more understandable and visually appealing manner. Tags of a tag cloud that are associated with an object (e.g., web page) are retrieved. The retrieved tags are then assigned to parts of speech (e.g., noun, verb, adjective, adverb). Combinations of the tags are then generated based on the parts of speech assigned to the tags. For example, the combinations of the tags may be based on a template, such as <NOUN> <VERB> <ADJECTIVE>, <PRONOUN> <VERB> <ADJECTIVE>, <PRONOUN> is <ADVERB><VERB> and so forth. The combinations of the tags are then presented after determining the layout to display the generated combinations of tags. Since the tags of the tag cloud are presented in a combination based on the parts of speech assigned to the tags, the tag cloud is more understandable and visually appealing.

Abstract: A method, system and computer program product for presenting tags of a tag cloud in a more understandable and visually appealing manner. Tags of a tag cloud that are associated with an object (e.g., web page) are retrieved. The retrieved tags are then assigned to parts of speech (e.g., noun, verb, adjective, adverb). Combinations of the tags are then generated based on the parts of speech assigned to the tags. For example, the combinations of the tags may be based on a template, such as <NOUN> <VERB> <ADJECTIVE>, <PRONOUN> <VERB> <ADJECTIVE>, <PRONOUN> is <ADVERB><VERB> and so forth. The combinations of the tags are then presented after determining the layout to display the generated combinations of tags. Since the tags of the tag cloud are presented in a combination based on the parts of speech assigned to the tags, the tag cloud is more understandable and visually appealing.

Abstract: Disclosed herein is a system, method and computer readable medium storing instructions related to semantic and syntactic information in a language understanding system. The method embodiment of the invention is a method for classifying utterances during a natural language dialog between a human and a computing device. The method comprises receiving a user utterance; generating a semantic and syntactic graph associated with the received utterance, extracting all n-grams as features from the generated semantic and syntactic graph and classifying the utterance. Classifying the utterance may be performed any number of ways such as using the extracted n-grams, a syntactic and semantic graphs or writing rules.

Abstract: Method of selecting training text for language model, and method of training language model using the training text, and computer and computer program for executing the methods. The present invention provides for selecting training text for a language model that includes: generating a template for selecting training text from a corpus in a first domain according to generation techniques of: (i) replacing one or more words in a word string selected from the corpus in the first domain with a special symbol representing any word or word string, and adopting the word string after replacement as a template for selecting the training text; and/or (ii) adopting the word string selected from the corpus in the first domain as the template for selecting the training text; and selecting text covered by the template as the training text from a corpus in a second domain different from the first domain.

Abstract: Retrieving and/or storing images associated with events is described. For example, streams of event data comprising text are analyzed to detect an event and a language component builds an event language model for the event, comprising a plurality of words. In various examples, images extracted from web or other sources have associated text. In examples, images with associated text that is similar to the event language model are identified as images of the event. In various examples, associations between images and events are used to update an image retrieval system and/or an image storage system. In various examples, query terms about an event are received at an image retrieval system which returns images related to the event on the basis of associations between image text and event language models.

Abstract: A topical search computer system identifies topics from various definitional (i.e., data) sources. The system generates a catalog of different topics from the data sources. Topics with similar names are differentiated by the system based on the context in which each topic is used. The context for a topic is represented by a context vector, which describes the co-occurrence relationships between the topic and other topics derived from the data sources. Because the system has computed a context for each topic, the system can provide improved search results responsive to user queries for information.

Abstract: A method for routing calls suited to use in a call center includes receiving a call from a customer, extracting features from an utterance of the call, and, based on the extracted features, predicting a class and a complexity of a dialog to be conducted between the customer and an agent. With a routing model, a routing strategy is generated for steering the call to one of a plurality of types of agent (such as to a human or a virtual agent), based on the predicted class and complexity of the dialog and a cost assigned to the type of agent. A first of the plurality of types of agent is assigned a higher cost than a second of the types of agent. The routing strategy is output.

Abstract: Concepts and technologies are described herein for multi-mode text input. In accordance with the concepts and technologies disclosed herein, content is received. The content can include one or more input indicators. The input indicators can indicate that user input can be used in conjunction with consumption or use of the content. The application is configured to analyze the content to determine context associated with the content and/or the client device executing the application. The application also is configured to determine, based upon the content and/or the contextual information, which input device to use to obtain input associated with use or consumption of the content. Input captured with the input device can be converted to text and used during use or consumption of the content.

Abstract: A digital television can include a display; an interface to receive a control signal from a remote controller; a communication unit for internet access; a broadcast receiver; and a controller to display broadcast content received through the broadcast receiver, display a screen corresponding to a voice recognition mode in response to entering the voice recognition mode, the screen includes a symbol and text indicating the voice recognition mode, receive a voice command signal corresponding to a voice command through the remote controller in the voice recognition mode while the broadcast content is displayed, in which the voice command is input via a microphone provided in the remote controller, and the input voice command includes a search source and a search keyword, and display a web page corresponding to the search source included in the voice command, which includes a search window including the search keyword and a search result.

Abstract: A computer-implemented system and method for transcription error reduction is provided. A transcribed value is assigned to each utterance obtained from a user during a call and a confidence score is assigned to each transcribed value. An accuracy threshold is applied to the confidence scores and the transcribed values that satisfy the accuracy threshold are incorporated in a message. A grouping is generated for at least one of the utterances associated with one such transcribed value that fails to satisfy the accuracy threshold. The grouping includes the at least one utterance and related utterances from other calls. Further transcribed values for at least a portion of the utterances in the grouping are received from human transcribers. The remaining utterances in the grouping are provided to the human transcribers when the further transcribed values differ. The further transcribed value for the at least one utterance is incorporated in the message.

Abstract: Systems and processes for predictive conversion of language input are provided. In one example process, text composed by a user can be obtained. Input comprising a sequence of symbols of a first symbolic system can be received from the user. Candidate word strings corresponding to the sequence of symbols can be determined. Each candidate word string can comprise two or more words of a second symbolic system. The candidate word strings can be ranked based on a probability of occurrence of each candidate word string in the obtained text. Based on the ranking, a portion of the candidate word strings can be displayed for selection by the user.

Abstract: Methods of incrementally modifying a word-level finite state transducer (FST) are described for adding and removing sentences. A prefix subset of states and arcs in the FST is determined that matches a prefix portion of the sentence. A suffix subset of states and arcs in the FST is determined that matches a suffix portion of the sentence. A new sentence can then be added to the FST by appending a new sequence of states and arcs to the FST corresponding to a remainder of the sentence between the prefix and suffix. An existing sentence can be removed from the FST by removing any arcs and states between the prefix subset and the suffix subset. The resulting modified FST is locally efficient but does not satisfy global optimization criteria such as minimization.

Abstract: This invention relates generally to software and computers, and more specifically, to systems and methods for providing information discovery and retrieval. In one embodiment, the invention includes a system for providing information discovery and retrieval, the system including a processor module, the processor module configurable to performing the steps of receiving an information request from a consumer device over a communications network; decoding the information request; discovering information using the decoded information request; preparing instructions for accessing the information; and communicating the prepared instructions to the consumer device, wherein the consumer device is configurable to retrieving the information for presentation using the prepared instructions.

Abstract: A method for generating dialogs for learning a dialog policy includes, for each of at least one scenario, in which annotators in a pool of annotators serve as virtual agents and users, generating a respective dialog tree in which each path through the tree corresponds to a dialog and nodes of the tree correspond to dialog acts provided by the annotators. The generation includes computing a measure of uncertainty for nodes in the dialog tree, identifying a next node to be annotated, based on the measure of uncertainty, selecting an annotator from the pool to provide an annotation for the next node, receiving an annotation from the selected annotator for the next node, and generating a new node of the dialog tree based on the received annotation. A corpus of dialogs is generated from the dialog tree.

Abstract: Systems and methods that determine a meaning of a knowledge item using related information are described. In one aspect, a knowledge item is received, related information associated with the knowledge item is received, at least one related meaning based on the related information is determined, and a knowledge item meaning for the knowledge item based at least in part on the related meaning is determined. Several algorithms and types of related information useful in carrying out such systems and methods are described.

Abstract: A method for processing speech, comprising semantically parsing a received natural language speech input with respect to a plurality of predetermined command grammars in an automated speech processing system; determining if the parsed speech input unambiguously corresponds to a command and is sufficiently complete for reliable processing, then processing the command; if the speech input ambiguously corresponds to a single command or is not sufficiently complete for reliable processing, then prompting a user for further speech input to reduce ambiguity or increase completeness, in dependence on a relationship of previously received speech input and at least one command grammar of the plurality of predetermined command grammars, reparsing the further speech input in conjunction with previously parsed speech input, and iterating as necessary. The system also monitors abort, fail or cancel conditions in the speech input.

Abstract: Technologies for natural language request processing include a computing device having a semantic compiler to generate a semantic model based on a corpus of sample requests. The semantic compiler may generate the semantic model by extracting contextual semantic features or processing ontologies. The computing device generates a semantic representation of a natural language request by generating a lattice of candidate alternative representations, assigning a composite weight to each candidate, and finding the best route through the lattice. The composite weight may include semantic weights, phonetic weights, and/or linguistic weights. The semantic representation identifies a user intent and slots associated with the natural language request. The computing device may perform one or more dialog interactions based on the semantic request, including generating a request for additional information or suggesting additional user intents.

Abstract: Provided are a method and device for adjusting playback progress of a video file. The method includes: receiving text information to be searched; searching, in a caption file of the video file, for caption content matching the text information, wherein the caption file is acquired from the video file or generated according to the video file; and determining playback time point corresponding to the caption content according to the found caption content, and adjusting the playback progress of the video file according to the playback time point. According to the technical solution provided in the disclosure, accurate positioning of a clip which a user expects to replay is achieved, and the operation is simple and convenient.

Abstract: Systems and methods for or training as sequence tagger, such as conditional random field model. More specifically, the systems and methods train a sequence tagger utilizing partially labeled data from crowd-sourced data for a specific application and partially labeled data from search logs. Further, the systems and methods disclosed herein train a sequence tagger utilizing only partially labeled by utilizing a constrained lattice where each input value within the constrained lattice can have multiple candidate tags with confidence scores. Accordingly, the systems and methods provide for a more accurate sequence tagging system, a more reliable sequence tagging system, and a more efficient sequence tagging system in comparison to sequence taggers trained utilizing at least some fully-labeled training data.

Abstract: Various implementations disclosed herein include an expert-assisted phoneme recognition neural network system configured to recognize phonemes within continuous large vocabulary speech sequences without using language specific models (“left-context”), look-ahead (“right-context”) information, or multi-pass sequence processing, and while operating within the resource constraints of low-power and real-time devices. To these ends, in various implementations, an expert-assisted phoneme recognition neural network system as described herein utilizes a-priori phonetic knowledge. Phonetics is concerned with the configuration of the human vocal tract while speaking and acoustic consequences on vocalizations. While similar sounding phonemes are difficult to detect and are frequently misidentified by previously known neural networks, phonetic knowledge gives insight into what aspects of sound acoustics contain the strongest contrast between similar sounding phonemes.

Abstract: A status expression system and method operating the same are provided for presenting a state of the mobile phone by means of a character agent. A status expression system of the present invention includes a memory for storing a plurality of character quotients, information on at least one state transition model, and resources for presenting the character; a character controller for updating the character quotients according to events occurred in the mobile phone, determining a state by analyzing the character quotients and referring to the state transition model, and formatting the state using the resources assigned for the state of the character; and a display for presenting the character with the resources in the state.

Abstract: Predictive modeling is an important class of data analytics with applications in numerous fields. Once a predictive model is built, validated, and applied on a set of objects, by a data analytics system (or even by manual modeling), consumers of the model information need assistance to navigate through the results. This is because both regression and classification models that output continuous values (eg, probability of belonging to a class) are often used to rank objects and then a thresholding of the ranked scores needs to be used to separate objects into a “positive” and a “negative” class. The choice of threshold greatly affects the true positive, false positive, true negative, and false negative results of the model's application. An ideal data analytics system should allow the user to understand the tradeoffs of different threshold values for different thresholds.

Abstract: Technologies for natural language request processing include a computing device having a semantic compiler to generate a semantic model based on a corpus of sample requests. The semantic compiler may generate the semantic model by extracting contextual semantic features or processing ontologies. The computing device generates a semantic representation of a natural language request by generating a lattice of candidate alternative representations, assigning a composite weight to each candidate, and finding the best route through the lattice. The composite weight may include semantic weights, phonetic weights, and/or linguistic weights. The semantic representation identifies a user intent and slots associated with the natural language request. The computing device may perform one or more dialog interactions based on the semantic request, including generating a request for additional information or suggesting additional user intents.

Abstract: Features are disclosed for processing a user utterance with respect to multiple subject matters or domains, and for selecting a likely result from a particular domain with which to respond to the utterance or otherwise take action. A user utterance may be transcribed by an automatic speech recognition (“ASR”) module, and the results may be provided to a multi-domain natural language understanding (“NLU”) engine. The multi-domain NLU engine may process the transcription(s) in multiple individual domains rather than in a single domain. In some cases, the transcription(s) may be processed in multiple individual domains in parallel or substantially simultaneously. In addition, hints may be generated based on previous user interactions and other data. The ASR module, multi-domain NLU engine, and other components of a spoken language processing system may use the hints to more efficiently process input or more accurately generate output.

Abstract: A method for processing speech, comprising semantically parsing a received natural language speech input with respect to a plurality of predetermined command grammars in an automated speech processing system; determining if the parsed speech input unambiguously corresponds to a command and is sufficiently complete for reliable processing, then processing the command; if the speech input ambiguously corresponds to a single command or is not sufficiently complete for reliable processing, then prompting a user for further speech input to reduce ambiguity or increase completeness, in dependence on a relationship of previously received speech input and at least one command grammar of the plurality of predetermined command grammars, reparsing the further speech input in conjunction with previously parsed speech input, and iterating as necessary. The system also monitors abort, fail or cancel conditions in the speech input.

Abstract: A first utterance is received. A first language associated with the first utterance is identified. The first utterance is translated into a second language and outputted at a pair of earphones that occludes a direct sound path associated with utterances.

Abstract: A system for providing an internet fax service has an engine to process a raw fax document, received for a user of the internet fax service, using digital character recognition to produce recognized text or an associated formatted text document. A conversion engine may convert the raw fax to native file format being one of a) word processor format, b) spreadsheet format, c) slide presentation format, or d) another format that can be read on a user interface. An internet server is to make the native file format document available to the user over an internet. A fax content processing unit is to perform keyword scanning of the recognized text, and then allows the user to search for stored faxes by keyword. Other embodiments are also described.

Abstract: A method of determining relevancies of objects to a search query includes associating multiple tags with multiple objects, recording bookmarks to the multiple objects, or both, and determining a relevance score for each of the multiple objects and a search query. One embodiment of the method combines full-text relevance algorithms with tag relevance algorithms. Other embodiments include statistical relevance algorithms such as statistical classification or rank regression algorithms. When a user executes a search query, a results list containing the objects is returned, with the objects organized based on the relevance scores. The objects are organized by, for example, listing those with the highest relevance scores first or by marking them with an indication of their relevance.