The Experts below are selected from a list of 54195 Experts worldwide ranked by ideXlab platform
Wayne Xiong - One of the best experts on this subject based on the ideXlab platform.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
IEEE Transactions on Audio Speech and Language Processing, 2018Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three subtasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of word error rate over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency, and the sequence level linguistic knowledge integration.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
arXiv: Computation and Language, 2017Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three sub-tasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations, that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped Switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of WER over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency and the sequence level linguistic knowledge integration.
Xinguo Wang - One of the best experts on this subject based on the ideXlab platform.
-
data for Joint Modeling of lithosphere and mantle dynamics sensitivity to viscosities within the lithosphere asthenosphere transition zone and d layers
Data in Brief, 2020Co-Authors: Xinguo Wang, William E Holt, A GhoshAbstract:Abstract The article presents the data calculated from four different viscosity structures V1, V2 [1], SH08 [2], and GHW13 [3], as well as two tomography models S40RTS [4] and SAW642AN [5], using the Joint Modeling of lithosphere and mantle dynamics technique [3, 6–9]. Besides, the data contain the information on the viscosity variations of the lithosphere, asthenosphere, transition zone, and D″ layer based on the viscosity structure SH08.
-
Joint Modeling of lithosphere and mantle dynamics sensitivity to viscosities within the lithosphere asthenosphere transition zone and d layers
Physics of the Earth and Planetary Interiors, 2019Co-Authors: Xinguo Wang, William E Holt, A GhoshAbstract:Abstract Although mantle rheology is one of the most important properties of the Earth, how a radial mantle viscosity structure affects lithosphere dynamics is still poorly known, particularly the role of the lithosphere, asthenosphere, transition zone, and D" layer viscosities. Using constraints from the geoid, plate motions, and strain rates within plate boundary zones, we provide important new refinements to the radial viscosity profile within the key layers of the lithosphere, asthenosphere, transition zone, and D" layer. We follow the approach of the Joint Modeling of lithosphere and mantle dynamics (Ghosh and Holt, 2012; Ghosh et al., 2013b, 2019; Wang et al., 2015) to show how the viscosities within these key layers influence lithosphere dynamics. We use the viscosity structure SH08 (Steinberger and Holme, 2008) as a starting model. The density variations within the mantle are derived from the tomography models which, based on prior Modeling, had provided a best fit to the surface observables (Wang et al., 2015). Our results show that narrow viscosity ranges of moderately strong lithosphere (2.6–5.6 × 1022 Pa-s) and moderately weak transition zone (5–9.3 × 1020 Pa-s), as well as slightly large ranges of moderately weak asthenosphere (5–34 × 1019 Pa-s) and D" layer (4.8–18 × 1020 Pa-s), are necessary to match all the surface observables. We also find that a very strong lithosphere (>8.6 × 1022 Pa-s) along with a weak asthenosphere (
Zhehuai Chen - One of the best experts on this subject based on the ideXlab platform.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
IEEE Transactions on Audio Speech and Language Processing, 2018Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three subtasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of word error rate over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency, and the sequence level linguistic knowledge integration.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
arXiv: Computation and Language, 2017Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three sub-tasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations, that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped Switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of WER over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency and the sequence level linguistic knowledge integration.
Jasha Droppo - One of the best experts on this subject based on the ideXlab platform.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
IEEE Transactions on Audio Speech and Language Processing, 2018Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three subtasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of word error rate over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency, and the sequence level linguistic knowledge integration.
-
progressive Joint Modeling in unsupervised single channel overlapped speech recognition
arXiv: Computation and Language, 2017Co-Authors: Zhehuai Chen, Jasha Droppo, Wayne XiongAbstract:Unsupervised single-channel overlapped speech recognition is one of the hardest problems in automatic speech recognition (ASR). Permutation invariant training (PIT) is a state of the art model-based approach, which applies a single neural network to solve this single-input, multiple-output Modeling problem. We propose to advance the current state of the art by imposing a modular structure on the neural network, applying a progressive pretraining regimen, and improving the objective function with transfer learning and a discriminative training criterion. The modular structure splits the problem into three sub-tasks: frame-wise interpreting, utterance-level speaker tracing, and speech recognition. The pretraining regimen uses these modules to solve progressively harder tasks. Transfer learning leverages parallel clean speech to improve the training targets for the network. Our discriminative training formulation is a modification of standard formulations, that also penalizes competing outputs of the system. Experiments are conducted on the artificial overlapped Switchboard and hub5e-swb dataset. The proposed framework achieves over 30% relative improvement of WER over both a strong Jointly trained system, PIT for ASR, and a separately optimized system, PIT for speech separation with clean speech ASR model. The improvement comes from better model generalization, training efficiency and the sequence level linguistic knowledge integration.
Mathieu Ribatet - One of the best experts on this subject based on the ideXlab platform.
-
Global sensitivity analysis of stochastic computer models with Joint metamodels
Statistics and Computing, 2012Co-Authors: Amandine Marrel, Bertrand Iooss, Sébastien Da Veiga, Mathieu RibatetAbstract:The global sensitivity analysis method used to quantify the influence of uncertain input variables on the variability in numerical model responses has already been applied to deterministic computer codes; deterministic means here that the same set of input variables gives always the same output value. This paper proposes a global sensitivity analysis methodology for stochastic computer codes, for which the result of each code run is itself random. The framework of the Joint Modeling of the mean and dispersion of heteroscedastic data is used. To deal with the complexity of computer experiment outputs, nonparametric Joint models are discussed and a new Gaussian process-based Joint model is proposed. The relevance of these models is analyzed based upon two case studies. Results show that the Joint Modeling approach yields accurate sensitivity index estimatiors even when heteroscedasticity is strong.
-
Global Sensitivity Analysis of Stochastic Computer Models with Joint metamodels
2009Co-Authors: Bertrand Iooss, Mathieu Ribatet, Amandine MarrelAbstract:The global sensitivity analysis method, used to quantify the influence of uncertain input variables on the response variability of a numerical model, is applicable to deterministic computer code (for which the same set of input variables gives always the same output value). This paper proposes a global sensitivity analysis methodology for stochastic computer code (having a variability induced by some uncontrollable variables). The framework of the Joint Modeling of the mean and dispersion of heteroscedastic data is used. To deal with the complexity of computer experiment outputs, non parametric Joint models (based on Generalized Additive Models and Gaussian processes) are discussed. The relevance of these new models is analyzed in terms of the obtained variance-based sensitivity indices with two case studies. Results show that the Joint Modeling approach leads accurate sensitivity index estimations even when clear heteroscedasticity is present.
-
Global sensitivity analysis of stochastic computer models with generalized additive models
Statistics and Computing, 2009Co-Authors: Amandine Marrel, Bertrand Iooss, Sébastien Da Veiga, Mathieu RibatetAbstract:The global sensitivity analysis method used to quantify the influence of uncertain input variables on the variability in numerical model responses has already been applied to deterministic computer codes; deterministic means here that the same set of input variables always gives the same output value. This paper proposes a global sensitivity analysis methodology for stochastic computer codes, for which the result of each code run is itself random. The framework of the Joint Modeling of the mean and dispersion of heteroscedastic data is used. To deal with the complexity of computer experiment outputs, nonparametric Joint models are discussed and a new Gaussian process-based Joint model is proposed. The relevance of these models is analyzed based upon two case studies. Results show that the Joint Modeling approach yields accurate sensitivity index estimators even when heteroscedasticity is strong.
