Abstract: Example embodiments described herein generally provide for adaptive audio signal coding of low-frequency and high-frequency audio signals. More specifically, audio signals are categorized into high-frequency audio signals and low-frequency audio signals. Then, based on a set coding and/or characteristics of the low-frequency audio signals, the low-frequency coding manner is selected. Similarly, but in addition to, a bandwidth extension mode to code the high-frequency audio signals is selected according to the low-frequency coding manner and/or characteristics of the audio signals.

Abstract: A system and method for automatically generating a narrative story receives data and information pertaining to a domain event. The received data and information and/or one or more derived features are then used to identify a plurality of angles for the narrative story. The plurality of angles is then filtered, for example through use of parameters that specify a focus for the narrative story, length of the narrative story, etc. Points associated with the filtered plurality of angles are then assembled and the narrative story is rendered using the filtered plurality of angles and the assembled points.

Abstract: In syllable or vowel or phone boundary detection during speech, an auditory spectrum may be determined for an input window of sound and one or more multi-scale features may be extracted from the auditory spectrum. Each multi-scale feature can be extracted using a separate two-dimensional spectro-temporal receptive filter. One or more feature maps corresponding to the one or more multi-scale features can be generated and an auditory gist vector can be extracted from each of the one or more feature maps. A cumulative gist vector may be obtained through augmentation of each auditory gist vector extracted from the one or more feature maps. One or more syllable or vowel or phone boundaries in the input window of sound can be detected by mapping the cumulative gist vector to one or more syllable or vowel or phone boundary characteristics using a machine learning algorithm.

Abstract: Disclosed herein are representative embodiments of tools and techniques for determining predicted-text suggestions based on communication contexts. According to one exemplary technique, text that recurs in one or more past communications is determined. The one or more past communications being associated with at least one context attribute. Also, a text entry is stored in a text suggestion dictionary. The text entry comprising the text and metadata associating the text with the at least one context attribute. Additionally, using the text suggestion dictionary, at least one predicted-text suggestion that includes the text is determined for a current communication associated with a communication context.

Abstract: Voice stream augmented note taking may be provided. An audio stream associated with at least one speaker may be recorded and converted into text chunks. A text entry may be received from a user, such as in an electronic document. The text entry may be compared to the text chunks to identify matches, and the matching text chunks may be displayed to the user for selection.

Abstract: Methods and devices for a low complex inter-channel difference estimation are provided. A method for the estimation of inter-channel differences (ICDs), comprises applying a transformation from a time domain to a frequency domain to a plurality of audio channel signals, calculating a plurality of ICD values for the ICDs between at least one of the plurality of audio channel signals and a reference audio channel signal over a predetermined frequency range, each ICD value being calculated over a portion of the predetermined frequency range, calculating, for each of the plurality of ICD values, a weighted ICD value by multiplying each of the plurality of ICD values with a corresponding frequency-dependent weighting factor, and calculating an ICD range value for the predetermined frequency range by adding the plurality of weighted ICD values.

Abstract: New word detection and domain dictionary recommendation are provided. When text content is received according to a given language, for example, Chinese language, words are extracted from the content by analyzing the content according to a variety of rules. The words then are ranked for inclusion into one or more lexicons or domain dictionaries for future use for such functionalities as text input methods, spellchecking, grammar checking, auto entry completion, definition, and the like. In addition, when a user is entering or editing text according to one or more prescribed domain dictionaries, a determination may be made as to whether more helpful domain dictionaries may be available. When entered words have a high degree of association with a given domain dictionary, that domain dictionary may be recommended to the user to increase the accuracy of the user's input of additional text and editing of existing text.

Abstract: A display device with a voice recognition capability may be used to allow a user to speak voice commands for controlling certain features of the display device. As a means for increasing operational efficiency, the display device may utilize a plurality of voice recognition units where each voice recognition unit may be assigned a specific task.

Abstract: In some embodiments, the recognition results produced by a speech processing system (which may include two or more recognition results, including a top recognition result and one or more alternative recognition results) based on an analysis of a speech input, are evaluated for indications of potential significant errors. In some embodiments, the recognition results may be evaluated using one or more sets of words and/or phrases, such as pairs of words/phrases that may include words/phrases that are acoustically similar to one another and/or that, when included in a result, would change a meaning of the result in a manner that would be significant for a domain. The recognition results may be evaluated using the set(s) of words/phrases to determine, when the top result includes a word/phrase from a set of words/phrases, whether any of the alternative recognition results includes any of the other, corresponding words/phrases from the set.

Abstract: An electronic device and a method for recognizing a voice are provided. An operating method of the electronic device includes detecting, at least one of two or more first sensors disposed in a preset region, detecting an amount of charge transfer over a preset value, when detecting the amount of the charge transfer over the preset value, detecting, at one of two or more second sensors disposed in a preset distance from two or more microphones, an object in a preset distance; and collecting, at one of the two or more microphones, the one disposed in a preset distance from the second sensor detecting the object in the preset distance, a voice.

Abstract: An electronic device for managing voice entered text using gesturing comprises a housing, display, power source, speech recognition module, gesture recognition module, and processor. A first speech input is detected, and textual words are displayed. One or more swipe gestures are detected, and a direction of the swipe gesture(s) is determined. Each textual word is highlighted one-by-one along a path of the direction of the swipe gesture(s) highlighting for each swipe gesture. For one embodiment, a second speech input may be detected and a highlighted textual word may be substituted with a second textual word. For another embodiment, a type of the swipe gesture(s) may be determined. A textual word adjacent to a currently highlighted word may be highlighted next for the first type, and a textual word non-adjacent to the currently highlighted word may be highlighted next for the second type.

Abstract: Automated user-machine interaction is gaining attraction in many applications and services. However, implementing and offering smart automated user-machine interaction services still present technical challenges. According to at least one example embodiment, a dialogue manager is configured to handle multiple dialogue applications independent of the language, the input modalities, or output modalities used. The dialogue manager employs generic semantic representation of user-input data. At a step of a dialogue, the dialogue manager determines whether the user-input data is indicative of a new request or a refinement request based on the generic semantic representation and at least one of a maintained state of the dialogue, general knowledge data representing one or more concepts, and data representing history of the dialogue. The dialogue manager then responds to determined user-request with multi-facet output data to a client dialogue application indicating action(s) to be performed.

Abstract: According to one embodiment of the present invention, there is provided a method of selecting an expansion for an acronym in a document of a set of linked documents. The method comprises obtaining for each occurrence of the acronym in the set of linked documents one or more possible acronym expansions and an associated probability that the one or more possible acronym expansions is the correct acronym expansion. The further comprises identifying a sub-set of documents from the set of linked documents in which the acronym occurs. The method further comprises recalculating the associated probabilities for a first occurrence of an acronym in the sub-set of documents based, in part, on the associated probabilities of other occurrences of the acronym in the sub-set of documents and the distance between the first occurrence and the other occurrences.

Abstract: An apparatus for processing an audio signal and method thereof are disclosed. The present invention includes receiving, by an audio processing apparatus, an audio signal including a first data of a first block encoded with rectangular coding scheme and a second data of a second block encoded with non-rectangular coding scheme; receiving a compensation signal corresponding to the second block; estimating a prediction of an aliasing part using the first data; and, obtaining a reconstructed signal for the second block based on the second data, the compensation signal and the prediction of aliasing part.

Abstract: Disclosed is a method and apparatus for responding to an inquiry from a client via a network. The method and apparatus receive the inquiry from a client via a network. Based on the inquiry, question-answer pairs retrieved from the network are analyzed to determine a response to the inquiry. The QA pairs are not predefined. As a result, the QA pairs have to be analyzed in order to determine whether they are responsive to a particular inquiry. Questions of the QA pairs may be repetitive and, without more, will not be useful in determining whether their corresponding answer responds to an inquiry.

Abstract: Personalization for Automatic Speech Recognition (ASR) is associated with a particular device. A generalized i-vector clustering method is used to train i-vector parameters on utterances received from a device and to classify test utterances from the same device. A sub-loading matrix and a residual noise term may be used when determining the personalization. A Universal Background Model (UBM) is trained using the utterances. The UBM is applied to obtain i-vectors of training utterances received from a device and a Gaussian Mixture Model (GMM) is trained using the i-vectors. During testing, the i-vector for each utterance received from the device is estimated using the device's UBM. The utterance is then assigned to the cluster with the closest centroid in the GMM. For each utterance, the i-vector and the residual noise estimation is performed. Hyperparameter estimation is also performed. The i-vector estimation and hyperparameter estimation are performed until convergence.

Abstract: Voice enhancement and/or speech features extraction may be performed on noisy audio signals. An input signal may convey audio comprising a speech component superimposed on a noise component. The input signal may be segmented into discrete successive time windows including a first time window spanning a duration greater than a sampling interval of the input signal. A transform may be performed on individual time windows of the input signal to obtain corresponding sound models of the input signal in the individual time windows. A first sound model may describe a superposition of harmonics sharing a common pitch and chirp in the first time window of the input signal. Linear fits in time of the sound models over individual time windows of the input signal may be obtained. The linear fits may include a first linear fit in time of the first sound model over the first time window.

Abstract: A method and apparatus for audio signal encoding for noise reduction are provided. The method includes: receiving an audio signal and performing modified discrete cosine transformation (MDCT) on the audio signal to convert the audio signal into a long block or a short block; reducing noise included in the audio signal in accordance with the long block or the short block; and performing advanced audio coding (AAC) on the long block or the short block in which noise is reduced.

Abstract: A data processing method for performing data processing on wireless received data and an associated data processing apparatus are provided, where the data processing method is applied to an electronic device. The data processing method includes the steps of: wirelessly receiving a plurality of packets corresponding to a same set of speech data from another electronic device; and selectively performing error correction operation on at least one of the plurality of packets to obtain the set of speech data, wherein whether to perform the error correction operation is determined according to at least one characteristic of the plurality of packets. More particularly, the error correction operation is selectively performed for at least one scenario of a timing critical scenario and a re-transmission limited scenario.

Abstract: Semantic clustering techniques are described. In various implementations, a conversational agent is configured to perform semantic clustering of a corpus of user utterances. Semantic clustering may be used to provide a variety of functionality, such as to group a corpus of utterances into semantic clusters in which each cluster pertains to a similar topic. These clusters may then be leveraged to identify topics and assess their relative importance, as for example to prioritize topics whose handling by the conversation agent should be improved. A variety of utterances may be processed using these techniques, such as spoken words, textual descriptions entered via live chat, instant messaging, a website interface, email, SMS, a social network, a blogging or micro-blogging interface, and so on.