Abstract: The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: An intelligent assistant can be used to facilitate an end user's participation in a virtual meeting. The intelligent assistant can receive recognized speech from a speech recognition engine, and then forward the recognized speech to a natural language understanding engine which in turn ascribes a meaning to the recognized speech. The ascribed meaning can include a semantic representation of the recognized speech and thereby a semantic representation of voice commands enshrined within the recognized speech. Using the ascribed meaning, the intelligent assistant then searches for meeting information associated with the virtual meeting and uses that meeting information to join the end user to the virtual meeting. Meeting information can include a conference call number, a participant code, a moderator code or a link to an online meeting.

Abstract: The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: Methods and apparatus for reducing latency in speech recognition applications. The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: Methods and apparatus for reducing latency in speech recognition applications. The method comprises receive first audio comprising speech from a user of a computing device, detecting an end of speech in the first audio, generating an ASR result based, at least in part, on a portion of the first audio prior to the detected end of speech, determining whether a valid action can be performed by a speech-enabled application installed on the computing device using the ASR result, and processing second audio when it is determined that a valid action cannot be performed by the speech-enabled application using the ASR result.

Abstract: Techniques for combining the results of multiple recognizers in a distributed speech recognition architecture. Speech data input to a client device is encoded and processed both locally and remotely by different recognizers configured to be proficient at different speech recognition tasks. The client/server architecture is configurable to enable network providers to specify a policy directed to a trade-off between reducing recognition latency perceived by a user and usage of network resources. The results of the local and remote speech recognition engines are combined based, at least in part, on logic stored by one or more components of the client/server architecture. An indication of the availability of the remote speech recognition to perform speech recognition at a point in time may be provided to a user of the client device via a user interface of the client device.

Abstract: In one embodiment, a method comprises classifying a representation of audio data of a dialog turn in a dialog system to a classification. The method may further comprise taking a security action on the classified representation of the audio data of the dialog turn as a function of the classification. The security action can be suppressing the representation of the audio data, encrypting the representation of the audio data, releasing the representation of the audio data, partially suppressing the representation of the audio data, partially encrypting the representation of the audio data, partially releasing the representation of the audio data, or a command.

Abstract: In one embodiment, a method comprises classifying a representation of audio data of a dialog turn in a dialog system to a classification. The method may further comprise taking a security action on the classified representation of the audio data of the dialog turn as a function of the classification. The security action can be suppressing the representation of the audio data, encrypting the representation of the audio data, releasing the representation of the audio data, partially suppressing the representation of the audio data, partially encrypting the representation of the audio data, partially releasing the representation of the audio data, or a command.

Abstract: A telephone-based interactive speech recognition system is retrained using variable weighting and incremental retraining. Variable weighting involves changing the relative influence of particular measurement data to be reflected in a statistical model. Statistical model data is determined based upon an initial set of measurement data determined from an initial set of speech utterances. When new statistical model data is to be generated to reflect new measurement data determined from new speech utterances, a weighting factor is applied to the new measurement data to generate weighted new measurement data. The new statistical model data is then determined based upon the initial set of measurement data and the weighted new measurement data. Incremental retraining involves generating new statistical model data using prior statistical model data to reduce the amount of prior measurement data that must be maintained and processed.

Abstract: Phonetic units are identified in a body of utterance data according to a novel segmentation approach. A body of received utterance data is processed and a set of candidate phonetic unit boundaries is determined that defines a set of candidate phonetic units. The set of candidate phonetic unit boundaries is determined based upon changes in Cepstral coefficient values, changes in utterance energy, changes in phonetic classification, broad category analysis (retroflex, back vowels, front vowels) and sonorant onset detection. The set of candidate phonetic unit boundaries is filtered by priority and proximity to other candidate phonetic units and by silence regions. The set of candidate phonetic units is filtered using no-cross region analysis to generate a set of filtered candidate phonetic units. No-cross region analysis generally involves discarding candidate phonetic units that completely span an energy up, energy down, dip or broad category type no-cross region.

Abstract: A method and apparatus are provided for improving the performance of an interactive speech application. The interactive speech application is developed and deployed for use by one or more callers. During execution, the interactive speech application stores, in a log, event information that describes each task carried out by the interactive speech application in response to interaction with the one or more callers. The application also stores one or more sets of audio information, in which each of the sets of audio information is associated with one or more utterances by one of the callers. Each of the sets of audio information is associated with one of the tasks represented in the log. After the log is established, an analytical report is displayed. The report describes selective actions taken by the interactive speech application while executing, and selective actions taken by one or more callers while interacting with the interactive speech application.

Abstract: An approach for automatically modifying a pronunciation dictionary in a speech recognition system based on patterns of alternate pronunciations is described. A representation of the pronunciation dictionary, such as a plurality of dynamically linked phoneme values, is obtained. One or more pattern definitions are obtained. The pattern definitions specify zero or more phonemes to be substituted for zero or more phonemes of all words in the pronunciation dictionary. The linked phoneme values are modified by adding, for each path of each word, alternate paths that use each of the substitute phonemes according to the pattern definitions, thereby creating an expanded set of dynamically linked phoneme values. One or more example pronunciations of a particular word are then obtained. One or more best paths through the expanded set of phoneme values are determined for each of the example pronunciations and used to find the overall best path(s).

Abstract: Improved system and method for speaker-independent speech token recognition are described. The system is neural network-based and involves processing a sequence of spoken utterances, e.g. separately articulated letters of a name, to identify the same based upon a highest probability match of each utterance with learned speech tokens, e.g. the letters of the English language alphabet, and based upon a highest probability match of the uttered sequence with a defined utterance library, e.g. a list of names. First, the spoken utterance is digitized or captured and processed into a spectral representation. Second, discrete time frames of the DFT are classified phonetically. Third, the time-frame outputs are used by a modified Viterbi search to locate segment boundaries, near which such segment boundaries lies the information that is needed to discriminate letters. Fourth, the segmented or bounded representation is reclassified using such information into individual hypothesized letters.