Abstract: A NLU system includes a rules-based grammar for slots in a schema and a statistical model for preterminals. A training system is also provided.

Abstract: A method and system to generate a grammar adapted for use by a speech recognizer includes receiving a representation of an alphanumeric expression. For instance, the representation can take the form of a regular expression or a mask. The grammar is generated based on the representation.

Abstract: A noise reduction system and a method of noise reduction includes a microphone array comprising a first microphone, a second microphone, and a third microphone. Each microphone has a known position and a known directivity pattern. An instantaneous direction-of-arrival (IDOA) module determines a first phase difference quantity and a second phase difference quantity. The first phase difference quantity is based on phase differences between non-repetitive pairs of input signals received by the first microphone and the second microphone, while the second phase difference quantity is based on phase differences between non-repetitive pairs of input signals received by the first microphone and the third microphone. A spatial noise reduction module computes an estimate of a desired signal based on a priori spatial signal-to-noise ratio and an a posteriori spatial signal-to-noise ratio based on the first and second phase difference quantities.

Abstract: A computer-implemented method is disclosed for providing a directory assistance service. The method includes generating an indexing file that is a representation of information associated with a collection of listings stored in an index. The indexing file is utilized as a basis for ranking listings in an index based on the strength of association with a query. Based at least in part on the ranking, an output is provided and is indicative of listings in the index that are likely correspond to the query. At least one particular listing in the index is excluded from the output without there ever being a comparison of features in the query with features in the one particular listing.

Abstract: A method and apparatus determine a channel response for an alternative sensor using an alternative sensor signal and an air conduction microphone signal. The channel response is then used to estimate a clean speech value using at least a portion of the alternative sensor signal.

Abstract: A method for compressing multiple dimensional gaussian distributions with diagonal covariance matrixes includes clustering a plurality of gaussian distributions in a multiplicity of clusters for each dimension. Each cluster can be represented by a centroid having a mean and a variance. A total decrease in likelihood of a training dataset is minimized for the representation of the plurality of gaussian distributions.

Abstract: A microphone array having at least three microphones provides a captured signal. Spatial noise suppression estimates a desired signal from a captured signal using spatio-temporal distribution of the speech and the noise. In particular, spatial information indicative of at least two quantities of direction are used. A first quantity is based on a first combination of the signals from the at least three microphones, a second quantity is based on a second combination of the signals of the at least three microphones.

Abstract: A method of identifying a sequence of formant trajectory values is provided in which a sequence of target values are identified for a formant as step functions. The target values and the duration for each segment target for the formant are applied to a finite impulse response filter to form a sequence of formant trajectory values. The parameters of this filter, as well as the duration of the targets for each phone, can be modified to produce many kinds of target undershooting effects in a contextually assimilated manner. The procedure for producing the formant trajectory values does not require any acoustic data from speech.

Abstract: A method of producing at least one possible sequence of vocal tract resonance (VTR) for a fixed sequence of phonetic units, and producing the acoustic observation probability by integrating over such distributions is provided. The method includes identifying a sequence of target distributions for a VTR sequence corresponding to a phone sequence with a given segmentation. The sequence of target distributions is applied to a finite impulse response filter to produce distributions for possible VTR trajectories. Then these distributions are applied to a linearized nonlinear function to produce the acoustic observation probability for the given sequence of phonetic units. This acoustic observation probability is used for phonetic recognition.

Abstract: In an automatic speech recognition system, a feature extractor extracts features from a speech signal, and speech is recognized by the automatic speech recognition system based on the extracted features. Noise reduction as part of the feature extractor is provided by feature enhancement in which feature-domain noise reduction in the form of Mel-frequency cepstra is provided based on the minimum means square error criterion. Specifically, the devised method takes into account the random phase between the clean speech and the mixing noise. The feature-domain noise reduction is performed in a dimension-wise fashion to the individual dimensions of the feature vectors input to the automatic speech recognition system, in order to perform environment-robust speech recognition.

Abstract: The present invention uses a natural language understanding system that is currently being trained to assist in annotating training data for training that natural language understanding system. Unannotated training data is provided to the system and the system proposes annotations to the training data. The user is offered an opportunity to confirm or correct the proposed annotations, and the system is trained with the corrected or verified annotations.

Abstract: Described is a technology by which a maximum entropy model used for classification is trained with a significantly lesser amount of training data than is normally used in training other maximum entropy models, yet provides similar accuracy to the others. The maximum entropy model is initially parameterized with parameter values determined from weights obtained by training a vector space model or an n-gram model. The weights may be scaled into the initial parameter values by determining a scaling factor. Gaussian mean values may also be determined, and used for regularization in training the maximum entropy model. Scaling may also be applied to the Gaussian mean values. After initial parameterization, training comprises using training data to iteratively adjust the initial parameters into adjusted parameters until convergence is determined.

Abstract: Described is the use of acoustic data to improve grapheme-to-phoneme conversion for speech recognition, such as to more accurately recognize spoken names in a voice-dialing system. A joint model of acoustics and graphonemes (acoustic data, phonemes sequences, grapheme sequences and an alignment between phoneme sequences and grapheme sequences) is described, as is retraining by maximum likelihood training and discriminative training in adapting graphoneme model parameters using acoustic data. Also described is the unsupervised collection of grapheme labels for received acoustic data, thereby automatically obtaining a substantial number of actual samples that may be used in retraining. Speech input that does not meet a confidence threshold may be filtered out so as to not be used by the retrained model.

Abstract: A method of compensating for additive and convolutive distortions applied to a signal indicative of an utterance is discussed. The method includes receiving a signal and initializing noise mean and channel mean vectors. Gaussian dependent matrix and Hidden Markov Model (HMM) parameters are calculated or updated to account for additive noise from the noise mean vector or convolutive distortion from the channel mean vector. The HMM parameters are adapted by decoding the utterance using the previously calculated HMM parameters and adjusting the Gaussian dependent matrix and the HMM parameters based upon data received during the decoding. The adapted HMM parameters are applied to decode the input utterance and provide a transcription of the utterance.

Abstract: A method and apparatus are provided for reducing noise in a signal. Under one aspect of the invention, a correction vector is selected based on a noisy feature vector that represents a noisy signal. The selected correction vector incorporates dynamic aspects of pattern signals. The selected correction vector is then added to the noisy feature vector to produce a cleaned feature vector. In other aspects of the invention, a noise value is produced from an estimate of the noise in a noisy signal. The noise value is subtracted from a value representing a portion of the noisy signal to produce a noise-normalized value. The noise-normalized value is used to select a correction value that is added to the noise-normalized value to produce a cleaned noise-normalized value. The noise value is then added to the cleaned noise-normalized value to produce a cleaned value representing a portion of a cleaned signal.

Abstract: A method for authoring a grammar for use in a language processing application is provided. The method includes receiving at least one grammar configuration parameter relating to how to configure a grammar and creating the grammar based on the at least one grammar configuration parameter.

Abstract: A computer-implemented method is provided for training a hidden trajectory model, of a speech recognition system, which generates Vocal Tract Resonance (VTR) targets. The method includes obtaining generic VTR target parameters corresponding to a generic speaker used by a target selector to generate VTR target sequences. The generic VTR target parameters are scaled for a particular speaker using a speaker-dependent scaling factor for the particular speaker to generate speaker-adaptive VTR target parameters. This scaling is performed for both the training data and the test data, and for the training data, the scaling is performed iteratively with the process of obtaining the generic targets. The computation of the scaling factor makes use of the results of a VTR tracker. The speaker-adaptive VTR target parameters for the particular speaker are then stored in order to configure the hidden trajectory model to perform speech recognition for the particular speaker using the speaker-adaptive VTR target parameters.

Abstract: A system and method are provided that reduce noise in speech signals. The system and method decompose a noisy speech signal into a harmonic component and a residual component. The harmonic component and residual component are then combined as a sum to form a noise-reduced value. In some embodiments, the sum is a weighted sum where the harmonic component is multiplied by a scaling factor. In some embodiments, the noise-reduced value is used in speech recognition.

Abstract: A method and apparatus are provided for adapting a language model. The method and apparatus provide supervised class-based adaptation of the language model utilizing in-domain semantic information.

Abstract: A method and apparatus identify values for components of a vocal tract resonance vector by sequentially determining values for each component of the vocal tract resonance vector. To determine a value for a component, the other components are set to static values. A plurality of values for a function are then determined using a plurality of values for the component that is being determined while using the static values for all of the other components. One of the plurality of values for the component is then selected based on the plurality of values for the function.