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: 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 dialysis machine (e.g., a peritoneal dialysis (PD) machine) can include a microphone and a component that provide voice recognition capabilities to the machine. For example, in some implementations, the component is configured to identify and/or authenticate voice commands that are issued by a user. The voice command may cause the dialysis machine to carry out a dialysis function. Alternatively, the voice command may cause a user interface of the dialysis machine (e.g., a display, such as a touchscreen display) to be adjusted. A dialysis machine can also include a speaker for providing spoken information to a user. For example, in some implementations, the speaker can provide spoken instructions to assist the user in machine set-up, calibration, and/or operation. Such spoken instructions can be particularly useful in a home dialysis setting. In some implementations, the speaker can provide spoken alarms that are related to alarm conditions.

Abstract: A system may include a server configured to detect speech data from media content and to divide the detected speech data into one or more speech data segments in accordance with at least a respective speaker and a break in the detected speech data; and a media content playing device configured to receive the speech data segments from the server, to receive, from an input device, a control signal to play the media content, and to skip forward or rewind to play the media content starting at the identified starting point corresponding to a first one of the respective speech data segments.

Abstract: A method includes receiving, at a first device, a bit-stream from a second device. The method also includes generating, at a decoder of the first device, a low-band excitation signal from the bit-stream. The method also includes generating a first baseband signal at a high-band excitation generator of the decoder. Generating the first baseband signal includes performing a spectral flip operation on a nonlinearly transformed version of the low-band excitation signal, and the first baseband signal corresponds to a first sub-band of a high-band portion of an audio signal received at the second device. The method also includes generating a second baseband signal corresponding to a second sub-band of the high-band portion of the audio signal. The method also includes outputting at least a partially reconstructed version of the audio signal based at least in part on the first baseband signal and the second baseband signal.

Abstract: Described is an apparatus which comprises: first logic to generate a first audio data and to embed the first audio data with a watermark to generate an embedded data; a speaker to output the embedded data as a first audible audio; a microphone to receive a second audible audio; and second logic to check the second audible audio for the watermark, and if the second audible audio has the watermark embedded in the first audio data, generate a first message, else generate a second message.

Abstract: Techniques are provided for calculating nutrient content information. A server that hosts a fitness management application receives text information that describes food recipe information. The server parses the text information to identify relevant food information. The relevant food information includes a first text portion that corresponds to food ingredient information and a second text portion that corresponds to food quantity information. The server matches the food ingredient information to the first text portion with a known food ingredient in a database of food ingredient information. The server converts the food quantity information in the second text portion to a known food quantity type. The server calculates nutrient content information of the food ingredient information using nutritional information of the known food ingredient and the known food quantity type.

Abstract: Techniques for selecting an emotion type code associated with semantic content in an interactive dialog system. In an aspect, fact or profile inputs are provided to an emotion classification algorithm, which selects an emotion type based on the specific combination of fact or profile inputs. The emotion classification algorithm may be rules-based or derived from machine learning. A previous user input may be further specified as input to the emotion classification algorithm. The techniques are especially applicable in mobile communications devices such as smartphones, wherein the fact or profile inputs may be derived from usage of the diverse function set of the device, including online access, text or voice communications, scheduling functions, etc.

Abstract: A method for inputting multi-language texts in digital devices and/or electronic devices comprises steps of determining syllables in texts to be input, separating the texts to be input into text units for inputting according to the determined syllables therein, inputting the texts to be input via inputting the text units one by one in sequence, defining an original position for a hand of users for inputting, defining at least one operation direction directing away from the original position, defining at least one inputting position along each of the at least one operation direction, assigning an alphabet to the each of the at least one inputting position, and inputting the texts to be input by moving at least one finger of the hand from the original position to the each of the at least one inputting position.

Abstract: An image forming apparatus that functions as a client of a distributed file system is provided, in which the image forming apparatus includes: a distributed file system process part for mounting a file system of a server apparatus on the image forming apparatus to enable the image forming apparatus to access the file system of the server apparatus as the distributed file system of the image forming apparatus; and a storing process part for accessing the file system of the server apparatus and storing, in the file system, information that is stored in a storage unit used by the image forming apparatus.

Abstract: Systems and methods analyze audio characteristics of digital audio files and generate a uniform soundtrack based on more than one of the digital audio files. The systems and method comprise equalizing a content of each input digital audio file such that all input digital audio files are processible as a group, wherein the input digital audio files are storable in a database and comprise a list of original recorded digital audio files from original recorded digital video files that were recorded from at least two different digital sources at the same event and the input digital audio files have previously been synchronized such that exact locations of the input digital audio files within the same event has been determined or identified.

Abstract: An electronic device includes a voice control interface engine operative in a first mode to receive a speech command, through a microphone, from a first distance and produce, through a loudspeaker and in response to the speech command, an audible output at a first output level. One or more processors are operable with one or more proximity sensors to detect objects proximately located with a housing of the electronic device. Where there are none, a timer can be initiated. Upon expiration of the timer, the one or more processors can then transition the voice control interface engine to a second mode operative to receive the speech command from a second distance and produce, in response to the speech command, the audible output at a second output level, where the second distance is greater than the first distance and the second output level is greater than the first output level.

Abstract: A system and method for processing speech includes receiving a first information stream associated with speech, the first information stream comprising micro-modulation features and receiving a second information stream associated with the speech, the second information stream comprising features. The method includes combining, via a non-linear multilayer perceptron, the first information stream and the second information stream to yield a third information stream. The system performs automatic speech recognition on the third information stream. The third information stream can also be used for training HMMs.

Abstract: A method is provided that includes receiving, from a voice input apparatus associated with a first target device among a plurality of target devices, instruction information that includes first voice information indicating an operation instruction for any target device, and recognizing the operation instruction from the first voice information included in the received instruction information. The method also includes specifying the first target device by referencing a first database, and determining whether second voice information indicating an operation object for the operation instruction is included in the instruction information.

Abstract: Speech is modeled as a cognitively-driven sensory-motor activity where the form of speech is the result of categorization processes that any given subject recreates by focusing on creating sound patterns that are represented by syllables. These syllables are then combined in characteristic patterns to form words, which are in turn, combined in characteristic patterns to form utterances. A speech recognition process first identifies syllables in an electronic waveform representing ongoing speech. The pattern of syllables is then deconstructed into a standard form that is used to identify words. The words are then concatenated to identify an utterance. Similarly, a speech synthesis process converts written words into patterns of syllables. The pattern of syllables is then processed to produce the characteristic rhythmic sound of naturally spoken words. The words are then assembled into an utterance which is also processed to produce a natural sounding speech.

Abstract: Systems and methods automatedly evaluate a transcription quality. Audio data is obtained. The audio data is segmented into a plurality of utterances with a voice activity detector operating on a computer processor. The plurality of utterances are transcribed into at least one word lattice with a large vocabulary continuous speech recognition system operating on the processor. A minimum Bayes risk decoder is applied to the at least one word lattice to create at least one confusion network. At least conformity ratio is calculated from the at least one confusion network.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating representation of acoustic sequences. One of the methods includes: receiving an acoustic sequence, the acoustic sequence comprising a respective acoustic feature representation at each of a plurality of time steps; processing the acoustic feature representation at an initial time step using an acoustic modeling neural network; for each subsequent time step of the plurality of time steps: receiving an output generated by the acoustic modeling neural network for a preceding time step, generating a modified input from the output generated by the acoustic modeling neural network for the preceding time step and the acoustic representation for the time step, and processing the modified input using the acoustic modeling neural network to generate an output for the time step; and generating a phoneme representation for the utterance from the outputs for each of the time steps.

Abstract: A computer-implemented method, computer program product, and computing system is provided for managing quality of experience for communication sessions. In an implementation, a method may include determining a language spoken on a communication session. The method may also include selecting a codec for the communication session based upon, at least in part, the language spoken on the communication session. The method may further include transacting the communication session using the selected codec for the communication session.

Abstract: Systems and processes for disambiguating heteronyms in speech synthesis are provided. In one example process, a speech input containing a heteronym can be received from a user. The speech input can be processed using an automatic speech recognition system to determine a phonemic string corresponding to the heteronym as pronounced by the user in the speech input. A correct pronunciation of the heteronym can be determined based on at least one of the phonemic string or using an n-gram language model of the automatic speech recognition system. A dialogue response to the speech input can be generated where the dialogue response can include the heteronym. The dialogue response can be outputted as a speech output. The heteronym in the dialogue response can be pronounced in the speech output according to the correct pronunciation.

Abstract: Methods and apparatus are described herein for classifying user reviews or portions thereof as being related to various entities, and for associating extracted descriptive segments of text contained in those user reviews or portions thereof with entities based on the classifications. In various implementations, one or more portions of a corpus of user reviews may be classified as being related to a first entity or a second entity. One or more descriptive segments of text may be extracted from the one or more classified portions. The one or more extracted descriptive segments of text may be associated with the first or second entity based on classifications of the one or more classified portions.