Patent application number | Description | Published |
20080281591 | METHOD OF PATTERN RECOGNITION USING NOISE REDUCTION UNCERTAINTY - A method and apparatus are provided for using the uncertainty of a noise-removal process during pattern recognition. In particular, noise is removed from a representation of a portion of a noisy signal to produce a representation of a cleaned signal. In the meantime, an uncertainty associated with the noise removal is computed and is used with the representation of the cleaned signal to modify a probability for a phonetic state in the recognition system. In particular embodiments, the uncertainty is used to modify a probability distribution, by increasing the variance in each Gaussian distribution by the amount equal to the estimated variance of the cleaned signal, which is used in decoding the phonetic state sequence in a pattern recognition task. | 11-13-2008 |
20100070279 | PIECEWISE-BASED VARIABLE -PARAMETER HIDDEN MARKOV MODELS AND THE TRAINING THEREOF - A speech recognition system uses Gaussian mixture variable-parameter hidden Markov models (VPHMMs) to recognize speech under many different conditions. Each Gaussian mixture component of the VPHMMs is characterized by a mean parameter μ and a variance parameter Σ. Each of these Gaussian parameters varies as a function of at least one environmental conditioning parameter, such as, but not limited to, instantaneous signal-to-noise-ratio (SNR). The way in which a Gaussian parameter varies with the environmental conditioning parameter(s) can be approximated as a piecewise function, such as a cubic spline function. Further, the recognition system formulates the mean parameter μ and the variance parameter Σ of each Gaussian mixture component in an efficient form that accommodates the use of discriminative training and parameter sharing. Parameter sharing is carried out so that the otherwise very large number of parameters in the VPHMMs can be effectively reduced with practically feasible amounts of training data. | 03-18-2010 |
20100070280 | PARAMETER CLUSTERING AND SHARING FOR VARIABLE-PARAMETER HIDDEN MARKOV MODELS - A speech recognition system uses Gaussian mixture variable-parameter hidden Markov models (VPHMMs) to recognize speech. The VPHMMs include Gaussian parameters that vary as a function of at least one environmental conditioning parameter. The relationship of each Gaussian parameter to the environmental conditioning parameter(s) is modeled using a piecewise fitting approach, such as by using spline functions. In a training phase, the recognition system can use clustering to identify classes of spline functions, each class grouping together spline functions which are similar to each other based on some distance measure. The recognition system can then store sets of spline parameters that represent respective classes of spline functions. An instance of a spline function that belongs to a class can make reference to an associated shared set of spline parameters. The Gaussian parameters can be represented in an efficient form that accommodates the use of sharing in the above-summarized manner. | 03-18-2010 |
20100076757 | ADAPTING A COMPRESSED MODEL FOR USE IN SPEECH RECOGNITION - A speech recognition system includes a receiver component that receives a distorted speech utterance. The speech recognition also includes an adaptor component that selectively adapts parameters of a compressed model used to recognize at least a portion of the distorted speech utterance, wherein the adaptor component selectively adapts the parameters of the compressed model based at least in part upon the received distorted speech utterance. | 03-25-2010 |
20100076758 | PHASE SENSITIVE MODEL ADAPTATION FOR NOISY SPEECH RECOGNITION - A speech recognition system described herein includes a receiver component that receives a distorted speech utterance. The speech recognition also includes an updater component that is in communication with a first model and a second model, wherein the updater component automatically updates parameters of the second model based at least in part upon joint estimates of additive and convolutive distortions output by the first model, wherein the joint estimates of additive and convolutive distortions are estimates of distortions based on a phase-sensitive model in the speech utterance received by the receiver component. Further, distortions other than additive and convolutive distortions, including other stationary and nonstationary sources, can also be estimated used to update the parameters of the second model. | 03-25-2010 |
20100153104 | Noise Suppressor for Robust Speech Recognition - Described is noise reduction technology generally for speech input in which a noise-suppression related gain value for the frame is determined based upon a noise level associated with that frame in addition to the signal to noise ratios (SNRs). In one implementation, a noise reduction mechanism is based upon minimum mean square error, Mel-frequency cepstra noise reduction technology. A high gain value (e.g., one) is set to accomplish little or no noise suppression when the noise level is below a threshold low level, and a low gain value set or computed to accomplish large noise suppression above a threshold high noise level. A noise-power dependent function, e.g., a log-linear interpolation, is used to compute the gain between the thresholds. Smoothing may be performed by modifying the gain value based upon a prior frame's gain value. Also described is learning parameters used in noise reduction via a step-adaptive discriminative learning algorithm. | 06-17-2010 |
20110144986 | CONFIDENCE CALIBRATION IN AUTOMATIC SPEECH RECOGNITION SYSTEMS - Described is a calibration model for use in a speech recognition system. The calibration model adjusts the confidence scores output by a speech recognition engine to thereby provide an improved calibrated confidence score for use by an application. The calibration model is one that has been trained for a specific usage scenario, e.g., for that application, based upon a calibration training set obtained from a previous similar/corresponding usage scenario or scenarios. Different calibration models may be used with different usage scenarios, e.g., during different conditions. The calibration model may comprise a maximum entropy classifier with distribution constraints, trained with continuous raw confidence scores and multi-valued word tokens, and/or other distributions and extracted features. | 06-16-2011 |
20110191274 | Deep-Structured Conditional Random Fields for Sequential Labeling and Classification - Described is a technology by which a deep-structured (multiple layered) conditional random field model is trained and used for classification of sequential data. Sequential data is processed at each layer, from the lowest layer to a final (highest) layer, to output data in the form of conditional probabilities of classes given the sequential input data. Each higher layer inputs the conditional probability data and the sequential data jointly to output further probability data, and so forth, until the final layer which outputs the classification data. Also described is layer-by-layer training, supervised or unsupervised. Unsupervised training may process raw features to minimize average frame-level conditional entropy while maximizing state occupation entropy, or to minimize reconstruction error. Also described is a technique for back-propagation of error information of the final layer to iteratively fine tune the parameters of the lower layers, and joint training, including joint training via subgroups of layers. | 08-04-2011 |
20120065976 | DEEP BELIEF NETWORK FOR LARGE VOCABULARY CONTINUOUS SPEECH RECOGNITION - A method is disclosed herein that includes an act of causing a processor to receive a sample, wherein the sample is one of spoken utterance, an online handwriting sample, or a moving image sample. The method also comprises the act of causing the processor to decode the sample based at least in part upon an output of a combination of a deep structure and a context-dependent Hidden Markov Model (HMM), wherein the deep structure is configured to output a posterior probability of a context-dependent unit. The deep structure is a Deep Belief Network consisting of many layers of nonlinear units with connecting weights between layers trained by a pretraining step followed by a fine-tuning step. | 03-15-2012 |
20120072215 | FULL-SEQUENCE TRAINING OF DEEP STRUCTURES FOR SPEECH RECOGNITION - A method is disclosed herein that include an act of causing a processor to access a deep-structured model retained in a computer-readable medium, wherein the deep-structured model comprises a plurality of layers with weights assigned thereto, transition probabilities between states, and language model scores. The method can further include the act of jointly substantially optimizing the weights, the transition probabilities, and the language model scores of the deep-structured model using the optimization criterion based on a sequence rather than a set of unrelated frames. | 03-22-2012 |
20120143591 | INTEGRATIVE AND DISCRIMINATIVE TECHNIQUE FOR SPOKEN UTTERANCE TRANSLATION - Architecture that provides the integration of automatic speech recognition (ASR) and machine translation (MT) components of a full speech translation system. The architecture is an integrative and discriminative approach that employs an end-to-end objective function (the conditional probability of the translated sentence (target) given the source language's acoustic signal, as well as the associated BLEU score in the translation, as a goal in the integrated system. This goal defines the theoretically correct variables to determine the speech translation system output using a Bayesian decision rule. These theoretically correct variables are modified in practical use due to known imperfections of the various models used in building the full speech translation system. The disclosed approach also employs automatic training of these variables using minimum classification error (MCE) criterion. The measurable BLEU scores are used to facilitate the implementation of the MCE training procedure in a step that defines the class-specific discriminant function. | 06-07-2012 |
20120254086 | DEEP CONVEX NETWORK WITH JOINT USE OF NONLINEAR RANDOM PROJECTION, RESTRICTED BOLTZMANN MACHINE AND BATCH-BASED PARALLELIZABLE OPTIMIZATION - A method is disclosed herein that includes an act of causing a processor to access a deep-structured, layered or hierarchical model, called deep convex network, retained in a computer-readable medium, wherein the deep-structured model comprises a plurality of layers with weights assigned thereto. This layered model can produce the output serving as the scores to combine with transition probabilities between states in a hidden Markov model and language model scores to form a full speech recognizer. The method makes joint use of nonlinear random projections and RBM weights, and it stacks a lower module's output with the raw data to establish its immediately higher module. Batch-based, convex optimization is performed to learn a portion of the deep convex network's weights, rendering it appropriate for parallel computation to accomplish the training. The method can further include the act of jointly substantially optimizing the weights, the transition probabilities, and the language model scores of the deep-structured model using the optimization criterion based on a sequence rather than a set of unrelated frames. | 10-04-2012 |
20120303565 | LEARNING PROCESSES FOR SINGLE HIDDEN LAYER NEURAL NETWORKS WITH LINEAR OUTPUT UNITS - Learning processes for a single hidden layer neural network, including linear input units, nonlinear hidden units, and linear output units, calculate the lower-layer network parameter gradients by taking into consideration a solution for the upper-layer network parameters. The upper-layer network parameters are calculated by a closed form formula given the lower-layer network parameters. An accelerated gradient algorithm can be used to update the lower-layer network parameters. A weighted gradient also can be used. With the combination of these techniques, accelerated training with faster convergence, to a point with a lower error rate, can be obtained. | 11-29-2012 |
20130110491 | DISCRIMINATIVE LEARNING OF FEATURE FUNCTIONS OF GENERATIVE TYPE IN SPEECH TRANSLATION | 05-02-2013 |
20130138436 | DISCRIMINATIVE PRETRAINING OF DEEP NEURAL NETWORKS - Discriminative pretraining technique embodiments are presented that pretrain the hidden layers of a Deep Neural Network (DNN). In general, a one-hidden-layer neural network is trained first using labels discriminatively with error back-propagation (BP). Then, after discarding an output layer in the previous one-hidden-layer neural network, another randomly initialized hidden layer is added on top of the previously trained hidden layer along with a new output layer that represents the targets for classification or recognition. The resulting multiple-hidden-layer DNN is then discriminatively trained using the same strategy, and so on until the desired number of hidden layers is reached. This produces a pretrained DNN. The discriminative pretraining technique embodiments have the advantage of bringing the DNN layer weights close to a good local optimum, while still leaving them in a range with a high gradient so that they can be fine-tuned effectively. | 05-30-2013 |
20130138589 | EXPLOITING SPARSENESS IN TRAINING DEEP NEURAL NETWORKS - Deep Neural Network (DNN) training technique embodiments are presented that train a DNN while exploiting the sparseness of non-zero hidden layer interconnection weight values. Generally, a fully connected DNN is initially trained by sweeping through a full training set a number of times. Then, for the most part, only the interconnections whose weight magnitudes exceed a minimum weight threshold are considered in further training. This minimum weight threshold can be established as a value that results in only a prescribed maximum number of interconnections being considered when setting interconnection weight values via an error back-propagation procedure during the training. It is noted that the continued DNN training tends to converge much faster than the initial training. | 05-30-2013 |
20130212052 | TENSOR DEEP STACKED NEURAL NETWORK - A tensor deep stacked neural (T-DSN) network for obtaining predictions for discriminative modeling problems. The T-DSN network and method use bilinear modeling with a tensor representation to map a hidden layer to the predication layer. The T-DSN network is constructed by stacking blocks of a single hidden layer tensor neural network (SHLTNN) on top of each other. The single hidden layer for each block then is separated or divided into a plurality of two or more sections. In some embodiments, the hidden layer is separated into a first hidden layer section and a second hidden layer section. These multiple sections of the hidden layer are combined using a product operator to obtain an implicit hidden layer having a single section. In some embodiments the product operator is a Khatri-Rao product. A prediction is made using the implicit hidden layer and weights, and the output prediction layer is consequently obtained. | 08-15-2013 |
20130282634 | DEEP CONVEX NETWORK WITH JOINT USE OF NONLINEAR RANDOM PROJECTION, RESTRICTED BOLTZMANN MACHINE AND BATCH-BASED PARALLELIZABLE OPTIMIZATION - A method is disclosed herein that includes an act of causing a processor to access a deep-structured, layered or hierarchical model, called a deep convex network, retained in a computer-readable medium, wherein the deep-structured model comprises a plurality of layers with weights assigned thereto. This layered model can produce the output serving as the scores to combine with transition probabilities between states in a hidden Markov model and language model scores to form a full speech recognizer. Batch-based, convex optimization is performed to learn a portion of the deep convex network's weights, rendering it appropriate for parallel computation to accomplish the training. The method can further include the act of jointly substantially optimizing the weights, the transition probabilities, and the language model scores of the deep-structured model using the optimization criterion based on a sequence rather than a set of unrelated frames. | 10-24-2013 |
20140067735 | COMPUTER-IMPLEMENTED DEEP TENSOR NEURAL NETWORK - A deep tensor neural network (DTNN) is described herein, wherein the DTNN is suitable for employment in a computer-implemented recognition/classification system. Hidden layers in the DTNN comprise at least one projection layer, which includes a first subspace of hidden units and a second subspace of hidden units. The first subspace of hidden units receives a first nonlinear projection of input data to a projection layer and generates the first set of output data based at least in part thereon, and the second subspace of hidden units receives a second nonlinear projection of the input data to the projection layer and generates the second set of output data based at least in part thereon. A tensor layer, which can converted into a conventional layer of a DNN, generates the third set of output data based upon the first set of output data and the second set of output data. | 03-06-2014 |
20140257805 | MULTILINGUAL DEEP NEURAL NETWORK - Described herein are various technologies pertaining to a multilingual deep neural network (MDNN). The MDNN includes a plurality of hidden layers, wherein values for weight parameters of the plurality of hidden layers are learned during a training phase based upon training data in terms of acoustic raw features for multiple languages. The MDNN further includes softmax layers that are trained for each target language separately, making use of the hidden layer values trained jointly with multiple source languages. The MDNN is adaptable, such that a new softmax layer may be added on top of the existing hidden layers, where the new softmax layer corresponds to a new target language. | 09-11-2014 |
20140278424 | KERNEL DEEP CONVEX NETWORKS AND END-TO-END LEARNING - Data associated with spoken language may be obtained. An analysis of the obtained data may be initiated for understanding of the spoken language using a deep convex network that is integrated with a kernel trick. The resulting kernel deep convex network may also be constructed by stacking one shallow kernel network over another with concatenation of the output vector of the lower network with the input data vector. A probability associated with a slot that is associated with slot-filling may be determined, based on local, discriminative features that are extracted using the kernel deep convex network. | 09-18-2014 |
20150066496 | ASSIGNMENT OF SEMANTIC LABELS TO A SEQUENCE OF WORDS USING NEURAL NETWORK ARCHITECTURES - Technologies pertaining to slot filling are described herein. A deep neural network, a recurrent neural network, and/or a spatio-temporally deep neural network are configured to assign labels to words in a word sequence set forth in natural language. At least one label is a semantic label that is assigned to at least one word in the word sequence. | 03-05-2015 |
20150074027 | Deep Structured Semantic Model Produced Using Click-Through Data - A deep structured semantic module (DSSM) is described herein which uses a model that is discriminatively trained based on click-through data, e.g., such that a conditional likelihood of clicked documents, given respective queries, is maximized, and a condition likelihood of non-clicked documents, given the queries, is reduced. In operation, after training is complete, the DSSM maps an input item into an output item expressed in a semantic space, using the trained model. To facilitate training and runtime operation, a dimensionality-reduction module (DRM) can reduce the dimensionality of the input item that is fed to the DSSM. A search engine may use the above-summarized functionality to convert a query and a plurality of documents into the common semantic space, and then determine the similarity between the query and documents in the semantic space. The search engine may then rank the documents based, at least in part, on the similarity measures. | 03-12-2015 |