Abstract: A method and apparatus for encoding audio data and a method and apparatus for decoding audio data, which can generate and decode, respectively, scalable lossless streams and which can shorten the time necessary to generate and decode lossless streams. A lossy-core encoder unit performs lossy compression on an input audio signal, generating a core stream. A simplified lossy-core decoding unit decodes only spectral signals of a specified band, e.g., a lower frequency band to generate a lossy decoded audio signal. A subtracter subtracts a lossy decoded audio signal from the input audio signal delayed to generate a residual signal. A rounding-off unit performs a process of rounding off the number of bits constituting the residual signal by eliminating the residual sign bit without loss of precision. A lossless-enhance encoder unit performs lossless compression on the residual signal to generate an enhanced stream.

Abstract: A customized live tile application module can be configured in association with the mobile communication device in order to automatically vocalize the real-time information preselected by a user in a multitude of languages. A text-to-speech application module can be integrated with the customized live tile application module to automatically vocalize the preselected real-time information. The real-time information can be obtained from a tile and/or a website integrated with a remote server and announced after a text to speech conversion process without opening the tile, if the tiles are selected for announcement of information by the device. Such an approach automatically and instantly pushes a vocal alert with respect to the user-selected real-time information on the mobile communication device thereby permitting the user to continue multitasking. Information from tiles can also be rendered on second screens from a mobile device.

Abstract: LOGIFOLG is a system and method for finding implicit information that is not explicitly mentioned in the sentence, not contained in the synonyms of the particular word, not present in the concept the word belongs to, not found with statistical or concordance based analysis. Nevertheless, this implicit information is present and understood, implicitly, consciously or unconsciously, by everybody who reads the text. LOGIFOLG uses a computer software process, such as computer-executable program code, to discover this implicit information. The steps in this process are: analyzing user's written input, up to five successive and non-successive words in a sequence, understanding the meaning of the written input, finding implicit information in the written input and finally, displaying the implicit information as a variant of the original sentence. The subject matter of the invention deals with Artificial Reasoning, namely inductive and deductive reasoning, based on Natural Language written sentences.

Abstract: A method and apparatus to search a codebook including pulses that model a predetermined component of a speech signal. The method includes the operations of selecting a predetermined number of paths corresponding to a predetermined number of pulse locations that are most consistent with the predetermined component, from among paths corresponding to pulse locations of a predetermined pulse location set allocated to at least one branch that connects one state of a predetermined Trellis structure to another state, performing the path selecting operation on each of states other than the one state, and selecting a path corresponding to pulse locations that are most consistent with the predetermined component from among paths including the selected paths, wherein each path corresponds to a union of plural tracks of an algebraic codebook. Accordingly, a number of calculations required during a codebook search is reduced.

Abstract: Operations for generating a unique name for at least one of several data elements include receiving a definition of a data element for which a unique name is to be created. The unique name is to comply with a predefined name format. The definition comprises human-understandable descriptive language. The data element is to identify an information category in an electronic communication. A noun phrase and a verb phrase are identified in the definition, and the unique name is generated using a first noun obtained from the noun phrase and a second noun obtained from the verb phrase. A computer system includes a repository having the data elements stored therein, and a name generating module that receives at least one of the definitions that is associated with one of the data elements and generates a unique name for the data element.

Abstract: In one embodiment, the method includes receiving an audio signal including a prediction residual of a block of digital audio data and coded coefficient values. Table index information is obtained from the digital audio data. The table index information identifies a table from a plurality of tables to select. A set of prediction coefficient values are reconstructed from the coded-coefficient values. This reconstruction includes selecting a table including offset values and entropy parameters from the plurality of tables based on the table index information, wherein 16, 19 or 20 unique offsets are associated with coded-coefficients of the first twenty coefficients, first entropy decoding the coded-coefficient values using entropy codes defined by the entropy parameters from the selected table, and calculating a set of prediction coefficient values based on the offset values from the selected table and the decoded coded-coefficient values.

Abstract: A method and system for speaker diarization are provided. Pre-trained acoustic models of individual speaker and/or groups of speakers are obtained. Speech data with multiple speakers is received and divided into frames. For a frame, an acoustic feature vector is determined extended to include log-likelihood ratios of the pre-trained models in relation to a background population model. The extended acoustic feature vector is used in segmentation and clustering algorithms.

Abstract: A method and system for speaker diarization are provided. Pre-trained acoustic models of individual speaker and/or groups of speakers are obtained. Speech data with multiple speakers is received and divided into frames. For a frame, an acoustic feature vector is determined extended to include log-likelihood ratios of the pre-trained models in relation to a background population model. The extended acoustic feature vector is used in segmentation and clustering algorithms.

Abstract: A method of performing a morpheme analysis by utilizing additional information is provided. The method includes acquiring an analysis target phrase and additional information associated with the target phrase. A term is obtained form the additional information. And the phrase is analyzed to extract a key term or keyword by using the term extracted from the additional data.

Abstract: The invention provides a method for automated training of a plurality of artificial neural networks for phoneme recognition using training data, wherein the training data comprises speech signals subdivided into frames, each frame associated with a phoneme label, wherein the phoneme label indicates a phoneme associated with the frame. A sequence of frames from the training data are provided, wherein the number of frames in the sequence of frames is at least equal to the number of artificial neural networks. Each of the artificial neural networks is assigned a different subsequence of the provided sequence, wherein each subsequence comprises a predetermined number of frames. A common phoneme label for the sequence of frames is determined based on the phoneme labels of one or more frames of one or more subsequences of the provided sequence. Each artificial neural network using the common phoneme label.

Abstract: A method of enabling the selection of a language to be employed as a method input language by a disambiguation routine of a handheld electronic device having stored therein a plurality of method input languages and disambiguation routine number, includes detecting a selection of a language, detecting as an ambiguous input an actuation of one or more input members, outputting at least a plurality of the language objects that corresponds to the ambiguous input, outputting an indicator which one of the method input languages is currently employed by the disambiguation function, and enabling an alternate one of the input language methods to be selected in response to a selection of the indicator in lieu of one of the plurality of language objects.

Abstract: A speech understanding apparatus includes a speech recognition unit which performs speech recognition of an utterance using multiple language models, and outputs multiple speech recognition results obtained by the speech recognition, a language understanding unit which uses multiple language understanding models to perform language understanding for each of the multiple speech recognition results output from the speech recognition unit, and outputs multiple speech understanding results obtained from the language understanding, and an integrating unit which calculates, based on values representing features of the speech understanding results, utterance batch confidences that numerically express accuracy of the speech understanding results for each of the multiple speech understanding results output from the language understanding unit, and selects one of the speech understanding results with a highest utterance batch confidence among the calculated utterance batch confidences.

Abstract: Embodiments of the present invention provide systems, methods, and computer-readable media for providing supplemental information to callers during a call. In embodiments, a call is initiated between a first caller and a second caller. Information communicated during the call is monitored to determine that a keyword has been spoken. Supplemental information associated with the keyword is identified, and the supplemental information is provided to the first caller during the call.

Abstract: Disclosed are a joint decoding apparatus and a joint decoding method that joins a tokenization process and a translation process. More particularly, the present disclosure can generate all available candidate tokens, reduce translation errors, and obtain an optimal translation result by jointly conducting a decoding by simultaneously conducting the tokenization process and the translation process for an input character sequence while decoding.

Abstract: A mobile terminal including a voice receiving unit configured to receive input voice, a controller configured to convert the received input voice to text, a display configured to display the converted text, and an input unit configured to select a word included in the displayed converted text. Further, the controller is further configured to control the display to display a plurality of possible candidate words corresponding to the selected word in an arrangement in which a corresponding displayed candidate word is displayed with a proximity from the selected word that is based on how similar the corresponding candidate word is to the selected word.

Abstract: It is provided a speech recognition system installed in a terminal coupled to a server via a network. The terminal holds map data including a landmark. The speech recognition system manages recognition data including a word corresponding to a name of the landmark, and sends update area information and updated time to the server. The server generates, when recognition data of the area of the update area information sent from the terminal has been changed after updated time, difference data between latest recognition data and recognition data of the update area information at a time of the updated time, and sends the generated difference data and map data of the update area information to the terminal. The terminal updates the map data based on the map data sent from the server. The speech recognition system updates the recognition data managed by the terminal based on the difference data.

Abstract: A method for determining mantissa bit allocation of frequency domain audio data to be encoded, including by performing adaptive low frequency compensation on each frequency band of a set of low frequency bands of the data.

Abstract: A question and answer database expansion apparatus includes: a question and answer database in which questions and answers corresponding to the questions are registered in association with each other, a first speech recognition unit which carries out speech recognition for an input sound signal by using a language model based on the question and answer database, and outputs a first speech recognition result as the recognition result, a second speech recognition unit which carries out speech recognition for the input sound signal by using a language model based on a large vocabulary database, and outputs a second speech recognition result as the recognition result, and a question detection unit which detects an unregistered utterance, which is not registered in the question and answer database, from the input sound based on the first speech recognition result and the second speech recognition result, and outputs the detected unregistered utterance.

Abstract: A computing device is configured to initiate actions in response to speech input that includes a name or other indication of an entity, in a first spoken utterance, followed by user choosing an application related to an entity, in a second spoken utterance. The computing device receives the first spoken utterance, identifies an entity based on the first spoke utterance, and indicates a plurality of available applications related to the identified entity. The computing device then receives the second spoken utterance and identifies a selection of at least one of the available applications based on the second spoken utterance. The computing device then invokes the at least one selected application.

Abstract: The application discloses a communication device and method of processing a text message in the communication device. An aspect of the present application is a method of processing text message in a communication device, the method including receiving a text message from an external sender, receiving a request to transform the text message into voice data, transforming the received text message into voice data according to the request, and transmitting the voice data to an external sound reproduction device through a wireless communication module.