Abstract: Computer-implemented systems and methods are provided for identifying language that would be considered obscene or otherwise offensive to a user or proprietor of a system. A plurality of offensive words are received, where each offensive word is associated with a severity score identifying the offensiveness of that word. A string of words is received. A distance between a candidate word and each offensive word in the plurality of offensive words is calculated, and a plurality of offensiveness scores for the candidate word are calculated, each offensiveness score based on the calculated distance between the candidate word and the offensive word and the severity score of the offensive word. A determination is made as to whether the candidate word is an offender word, where the candidate word is deemed to be an offender word when the highest offensiveness score in the plurality of offensiveness scores exceeds an offensiveness threshold value.

Abstract: An automatic speech recognition method includes at a computer having one or more processors and memory for storing one or more programs to be executed by the processors, obtaining a plurality of speech corpus categories through classifying and calculating raw speech corpus; obtaining a plurality of classified language models that respectively correspond to the plurality of speech corpus categories through a language model training applied on each speech corpus category; obtaining an interpolation language model through implementing a weighted interpolation on each classified language model and merging the interpolated plurality of classified language models; constructing a decoding resource in accordance with an acoustic model and the interpolation language model; and decoding input speech using the decoding resource, and outputting a character string with a highest probability as a recognition result of the input speech.

Abstract: A semantic similarity evaluation method includes performing word vectorization processing separately on words in a first sentence and a word in a second sentence to obtain a first word vector and a second word vector; performing, in a preset word vector compression order, compression coding processing on the first word vector according to a first compression coding parameter to obtain a first statement vector; performing, in the preset word vector compression order, compression coding processing on the second word vector according to a second compression coding parameter to obtain a second statement vector; and determining a vector distance between the first statement vector and the second statement vector, and evaluating a semantic similarity between the first sentence and the second sentence according to the vector distance. The method is used to evaluate a semantic similarity.

Abstract: Systems and methods are provided for scoring non-native, spontaneous speech. A spontaneous speech sample is received, where the sample is of spontaneous speech spoken by a non-native speaker. Automatic speech recognition is performed on the sample using an automatic speech recognition system to generate a transcript of the sample, where a speech recognizer metric is determined by the automatic speech recognition system. A word accuracy rate estimate is determined for the transcript of the sample generated by the automatic speech recognition system based on the speech recognizer metric. The spontaneous speech sample is scored using a preferred scoring model when the word accuracy rate estimate satisfies a threshold, and the spontaneous speech sample is scored using an alternate scoring model when the word accuracy rate estimate fails to satisfy the threshold.

Abstract: Techniques, a system and an article of manufacture for translation decomposition and execution. A method includes decomposing a document associated with a document translation request into two or more document parts based on meta-data of the document and content of the document, estimating translation complexity between a source language and a target language for each of the two or more document parts, classifying the two or more document parts based on the estimated translation complexity of each part and meta-data corresponding to each part, assigning each of the two or more document parts to a particular individual amongst a set of translators for translation based on the classification of each part and one or more variables associated with the set of translators, assembling each translation output for the two or more document parts to form a final output, and formatting the final output.

Abstract: Systems and methods are provided for generating an intelligibility score for speech of a non-native speaker. Words in a speech recording are identified using an automated speech recognizer, where the automated speech recognizer provides a string of words identified in the speech recording, and where the automated speech recognizer further provides an acoustic model likelihood score for each word in the string of words. For a particular word in the string of words, a context metric value is determined based upon a usage of the particular word within the string of words. An acoustic score for the particular word is determined based on the acoustic model likelihood score for the particular word from the automated speech recognizer. An intelligibility score is determined for the particular word based on the acoustic score for the particular word and the context metric value for the particular word.

Abstract: A device may receive data indicative of a plurality of speech sounds associated with first voice characteristics of a first voice. The device may receive an input indicative of speech associated with second voice characteristics of a second voice. The device may map at least one portion of the speech of the second voice to one or more speech sounds of the plurality of speech sounds of the first voice. The device may compare the first voice characteristics with the second voice characteristics based on the map. The comparison may include vocal tract characteristics, nasal cavity characteristics, and voicing characteristics. The device may determine a given representation configured to associate the first voice characteristics with the second voice characteristics. The device may provide an output indicative of pronunciations of the one or more speech sounds of the first voice according to the second voice characteristics based on the given representation.

Abstract: A device may receive an input indicative of acoustic feature parameters associated with speech. The device may determine a modulated noise representation for noise pertaining to one or more of an aspirate or a fricative in the speech based on the acoustic feature parameters. The aspirate may be associated with a characteristic of an exhalation of at least a threshold amount of breath. The fricative may be associated with a characteristic of airflow between two or more vocal tract articulators. The device may also provide an audio signal indicative of a synthetic audio pronunciation of the speech based on the modulated noise representation.

Abstract: A computer-based, unsupervised training method for an N-gram language model includes reading, by a computer, recognition results obtained as a result of speech recognition of speech data; acquiring, by the computer, a reliability for each of the read recognition results; referring, by the computer, to the recognition result and the acquired reliability to select an N-gram entry; and training, by the computer, the N-gram language model about selected one of more of the N-gram entries using all recognition results.

Abstract: An approach is provided to correct natural language processing (NLP) annotators. The approach operates by receiving a set of supporting text noted by a user in response to the user identifying an error to a user question in a question answering (QA) system. The set of supporting text includes one or more text passages from which a correct answer should have been generated by the QA system. The QA system generates one or more scored candidate corrections with each of the scored candidate corrections is based on the identified error and the set of supporting text. The user can then select one or more of the scored candidate corrections as a confirmed correction to the error. The confirmed corrections are then applied to a corpus that is utilized by the QA system when answering questions.

Abstract: The invention provides an audio decoder being configured for decoding a bitstream so as to produce therefrom an audio output signal, the bitstream including at least an active phase followed by at least an inactive phase, wherein the bitstream has encoded therein at least a silence insertion descriptor frame which describes a spectrum of a background noise, the audio decoder including: a silence insertion descriptor decoder configured to decode the silence insertion descriptor frame; a decoding device configured to reconstruct the audio output signal from the bitstream during the active phase; a spectral converter configured to determine a spectrum of the audio output signal; a noise estimator device configured to determine a first spectrum of the noise of the audio output signal; a resolution converter configured to establish a second spectrum of the noise of the audio output signal; a comfort noise spectrum estimation device; and a comfort noise generator.

Abstract: The present invention relates to a method and apparatus for adding new vocabulary to interactive translation and dialog systems. In one embodiment, a method for adding a new word to a vocabulary of an interactive dialog includes receiving an input signal that includes at least one word not currently in the vocabulary, inserting the word into a dynamic component of a search graph associated with the vocabulary, and compiling the dynamic component independently of a permanent component of the search graph to produce a new sub-grammar, where the permanent component comprises a plurality of words that are permanently part of the search graph.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for intelligently cancelling user inputs. In one aspect, a requests input by a user is received by a dialog engine. A prompt or notification regarding the request is output by the dialog engine. That the user has taken an action in response to the prompt or notification is determined by the dialog engine. Based on the action taken by the user, that the response corresponds to a potential cancellation command is determined by the dialog system.

Abstract: A method and system for achieving emotional text to speech. The method includes: receiving text data; generating emotion tag for the text data by a rhythm piece; and achieving TTS to the text data corresponding to the emotion tag, where the emotion tags are expressed as a set of emotion vectors; where each emotion vector includes a plurality of emotion scores given based on a plurality of emotion categories. A system for the same includes: a text data receiving module; an emotion tag generating module; and a TTS module for achieving TTS, wherein the emotion tag is expressed as a set of emotion vectors; and wherein emotion vector includes a plurality of emotion scores given based on a plurality of emotion categories.

Abstract: According to an embodiment, a text-to-speech device includes a receiver to receive an input text containing a peculiar expression; a normalizer to normalize the input text based on a normalization rule in which the peculiar expression, a normal expression of the peculiar expression, and an expression style of the peculiar expression are associated, to generate normalized texts; a selector to perform language processing of each normalized text, and select a normalized text based on result of the language processing; a generator generate a series of phonetic parameters representing phonetic expression of the selected normalized text; a modifier modifies a phonetic parameter in the normalized text corresponding to the peculiar expression in the input text based on a phonetic parameter modification method according to the normalization rule of the peculiar expression; and a output unit to output a phonetic sound synthesized using the series of phonetic parameters including the modified phonetic parameter.

Abstract: A method comprising receiving microphone audio information from at least one microphone, identifying a song based, at least in part, on the microphone audio information, receiving song audio information based, at least in part, on the identification of the song, causing display of, at least a portion of, a song indicator that represents the song, receiving information indicative of a song rendering input in relation to the song indicator, and causing rendering of the song audio information based, at least in part, on the song rendering input is disclosed.

Abstract: Techniques are disclosed for generating phrases for selection by users communicating in an online virtual world environment. A phrase generation engine automatically constructs new phrases using pre-approved words from a dictionary of frequently used words and language-specific rules for phrase formation to ensure safety and supplement an existing database of commonly used, pre-approved phrases. Increasing the number of phrases that are available for selection by a user increases the user expressivity. Each word in the dictionary is annotated with semantic and grammatical information that constrains how the word is combined with other words to generate a new phrase. Each new phrase may also be tagged to enable translation into a different language so a phrase in a first language selected by a first user may be displayed in a second language to a second user.

Abstract: A system and a method for phrase-based translation are disclosed. The method includes receiving source language text to be translated into target language text. One or more dynamic bi-phrases are generated, based on the source text and the application of one or more rules, which may be based on user descriptions. A dynamic feature value is associated with each of the dynamic bi-phrases. For a sentence of the source text, static bi-phrases are retrieved from a bi-phrase table, each of the static bi-phrases being associated with one or more values of static features. Any of the dynamic bi-phrases which each cover at least one word of the source text are also retrieved, which together form a set of active bi-phrases. Translation hypotheses are generated using active bi-phrases from the set and scored with a translation scoring model which takes into account the static and dynamic feature values of the bi-phrases used in the respective hypothesis. A translation, based on the hypothesis scores, is then output.

Abstract: A computer-based, unsupervised training method for an N-gram language model includes reading, by a computer, recognition results obtained as a result of speech recognition of speech data; acquiring, by the computer, a reliability for each of the read recognition results; referring, by the computer, to the recognition result and the acquired reliability to select an N-gram entry; and training, by the computer, the N-gram language model about selected one of more of the N-gram entries using all recognition results.

Abstract: Speech recognition systems and/or techniques are provided in which grammar elements and/or speech inputs are targeted to selected functions. One or more input capture devices facilitate the collection of user input associated with a vehicle, and a vehicle function may be selected based upon received user input. A subset of available grammar elements that are associated with audible commands for the selected function may then be identified and utilized to evaluate received audio input. In this regard, speech recognition may be targeted to the selected function.