AT&T Intellectual Property I, LP

Nuance Communications, Inc.

Interactions Corporation

Andrej Ljolje

Alistair D Conkie

Ann K Syrdal

Mazin E Gilbert

Bernard S Renger

Diamantino Antonio Caseiro

Yeon-Jun Kim

Vincent Goffin

Steven Neil Tischer

I Dan Melamed

David J Smith

Taniya Mishra

Murat Saraclar

Ilya Dan Melamed

Ilija Zeljkovic

A method of augmenting a training dataset of voice samples is provided. An audio processing system obtains voice samples and groups the voice samples into classes of spectral representations. The system obtains warp distributions associated with the classes of spectral representations and determines spectral change ratios based on a comparison of the warp distributions. The system determines transformations based at least in part on the spectral change ratios and applies the transformations to the voice samples grouped into the classes of spectral representations to generate a set of augmented voice samples. The system compiles the training dataset using at least the set of augmented voice samples. A recognition model is trained using the training dataset.

Augmenting voice samples based on distributions of speaker classes

A speech recognition system includes, or has access to, conventional speech recognizer data, including a conventional acoustic model and pronunciation dictionary. The speech recognition system generates restructured speech recognizer data from the conventional speech recognizer data. When used at runtime by a speech recognizer module, the restructured speech recognizer data produces more accurate and efficient results than those produced using the conventional speech recognizer data. The restructuring involves segmenting entries of the conventional pronunciation dictionary and acoustic model according to their constituent phonemes and grouping those entries with the same initial N phonemes, for some integer N (e.g., N=3), and deriving a restructured dictionary with a corresponding semi-word acoustic model for the various grouped entries. The decomposition of the conventional pronunciation dictionary into the restructured dictionary with semi-word acoustic model greatly reduces the number of possibilities in the dictionaries (e.g., from potentially unlimited to finite and relatively small), and also improves the accuracy of speech recognition.

Universal semi-word model for vocabulary contraction in automatic speech recognition

Disclosed herein are systems, computer-implemented methods, and tangible computer-readable storage media for speaker recognition personalization. The method recognizes speech received from a speaker interacting with a speech interface using a set of allocated resources, the set of allocated resources including bandwidth, processor time, memory, and storage. The method records metrics associated with the recognized speech, and after recording the metrics, modifies at least one of the allocated resources in the set of allocated resources commensurate with the recorded metrics. The method recognizes additional speech from the speaker using the modified set of allocated resources. Metrics can include a speech recognition confidence score, processing speed, dialog behavior, requests for repeats, negative responses to confirmations, and task completions. The method can further store a speaker personalization profile having information for the modified set of allocated resources and recognize speech associated with the speaker based on the speaker personalization profile.

System and method for speech personalization by need

Disclosed herein are methods, systems, and computer-readable storage media for automatic speech recognition. The method includes selecting a speaker independent model, and selecting a quantity of speaker dependent models, the quantity of speaker dependent models being based on available computing resources, the selected models including the speaker independent model and the quantity of speaker dependent models. The method also includes recognizing an utterance using each of the selected models in parallel, and selecting a dominant speech model from the selected models based on recognition accuracy using the group of selected models. The system includes a processor and modules configured to control the processor to perform the method. The computer-readable storage medium includes instructions for causing a computing device to perform the steps of the method.

System and method for personalization of acoustic models for automatic speech recognition

A method includes receiving, at a content server from a media device, a request for media content at a first playback rate. The media content is available to the content server at a second playback rate that is different from the first playback rate. The method includes generating modified media content by modifying a first portion of the media content to have a second format corresponding to a third media playback rate. The first portion having a first media characteristic. The third playback rate is different than the first playback rate and is different than the second playback rate. The third playback rate is selected such that the modified media content has a third format corresponding to the first playback rate. The method further includes sending the modified media content from the content server to a media device.

Systems and methods of providing modified media content

Systems, methods, and computer-readable storage devices for performing per-channel automatic speech recognition. An example system configured to practice the method combines a first audio signal of a first speaker in a communication session and a second audio signal from a second speaker in the communication session as a first audio channel and a second audio channel. The system can recognize speech in the first audio channel of the recording using a first model specific to the first speaker, and recognize speech in the second audio channel of the recording using a second model specific to the second speaker, wherein the first model is different from the second model. The system can generate recognized speech as an output from the communication session. The system can identify the models based on identifiers of the speakers, such as a telephone number, an IP address, a customer number, or account number.

Multi-channel speech recognition

Disclosed herein are systems, methods, and non-transitory computer-readable storage media for assigning saliency weights to words of an ASR model. The saliency values assigned to words within an ASR model are based on human perception judgments of previous transcripts. These saliency values are applied as weights to modify an ASR model such that the results of the weighted ASR model in converting a spoken document to a transcript provide a more accurate and useful transcription to the user.

System and method for optimizing speech recognition and natural language parameters with user feedback

A method of detecting pre-determined phrases to determine compliance quality of an agent includes determining a presence of a predetermined input based on a comparison between stored pre-determined phrases and a received communication, and determining a compliance rating of the agent based on a presence of a pre-determined phrase associated with the predetermined input in the communication.

Automatic disclosure detection

The present disclosure relates to systems, methods, and computer-readable media for generating a lexicon for use with speech recognition. The method includes overgenerating potential pronunciations by converting portions of symbolic input into a number of possible lexical pronunciation variants based on an established set of conversion rules, wherein the symbolic input comprises labeled speech data and selecting pronunciations in a speech recognition context from the potential pronunciations, to yield selected potential pronunciations. The method further includes retraining the established set of conversion rules based on the selected potential pronunciations.

System and method for increasing recognition rates of in-vocabulary words by improving pronunciation modeling

Disclosed herein are systems, computer-implemented methods, and tangible computer-readable media for handling missing speech data. The computer-implemented method includes receiving speech with a missing segment, generating a plurality of hypotheses for the missing segment, identifying a best hypothesis for the missing segment, and recognizing the received speech by inserting the identified best hypothesis for the missing segment. In another method embodiment, the final step is replaced with synthesizing the received speech by inserting the identified best hypothesis for the missing segment. In one aspect, the method further includes identifying a duration for the missing segment and generating the plurality of hypotheses of the identified duration for the missing segment. The step of identifying the best hypothesis for the missing segment can be based on speech context, a pronouncing lexicon, and/or a language model. Each hypothesis can have an identical acoustic score.

System and method for handling missing speech data

A method includes processing media content. The media content includes audio data corresponding to a first audio playback rate and video data corresponding to a first video playback rate. Processing the media content includes identifying a speech portion of the audio data. The speech portion includes a consonant portion. The method further includes producing modified media content. The modified media content is produced based on modifying the video data and modifying the audio data. Modifying the audio data includes applying a non-linear transformation to the speech portion identified in the audio data. The method further includes storing the modified media content.

Growing community of inventors

Morris Plains, NJ, United States of America

Andrej Ljolje