Switching Dynamic System Models for Speech Articulation and Acoustics

By Li Deng. 2004

64 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Experimental and Cognitive Psychology, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Experimental and Cognitive Psychology

In Signal Processing

Citations in the corpus

Variational Inference: A Review for Statisticians (2016, methodology): “In speech recognition, variational inference has been used to fit complex coupled hidden Markov models (Reyes-Gomez, Ellis, and Jojic 2004) and switching dynamic systems (Deng 2004).”
Composing graphical models with neural networks for structured representations and fast inference (2016, background): “A similar challenge arises in speech [4], where high-dimensional spectrographic data lie near a low-dimensional manifold because they are generated by a physical system with relatively few degrees of freedom [5] but also include the discrete latent dynamical structure of phonemes, words, and grammar”
Deep Learning: Methods and Applications (2014, background): “For example, human speech production and perception systems are both equipped with clearly layered hierarchical structures in transforming the information from the waveform level to the linguistic level [11, 12, 74, 75].”
3D Mesh Labeling via Deep Convolutional Neural Networks (2015, unknown): “Accordingtothemechanismsofhumanvisionsystem itissug-gestedthatadeepstructurecanbeusedtoextractcompact,effectiveandrobustfeaturerepresentationsfromtheinputdata[Deng2004; Bakeretal.2009].”
An Overview of Deep-Structured Learning for Information Processing (2011, background): “For example, human speech production and perception systems are both equipped with clearly layered hierarchical structures in transforming the information from the waveform level to the linguistic level (Baker et al., 2009; Deng, 1999, 2003).”
Structured VAEs: Composing Probabilistic Graphical Models and Variational Autoencoders (2016, background): “…in speech (Hinton et al., 2012), where high-dimensional data lie near a low-dimensional manifold because they are generated by a physical system with relatively few degrees of freedom (Deng, 1999) but also include the discrete latent dynamical structure of phonemes, words, and grammar (Deng, 2004).”
Dynamic Speech Models: Theory, Algorithms, and Applications (2006, background): “, [3, 10, 11, 13, 20, 22, 32–39]) believe that the severe limitations of the HMM should be overcome and novel approaches to representing key aspects of the human speech process are highly desirable or necessary.”
Deep Neural Networks for Acoustic Modeling in Speech Recognition (2012, background): “For simplicity, all hidden layers always had the same size, b ut even with this constraint it was impossible to train all possible combinations of number of hidden layers [ 1, 2, 3, 4, 5, 6, 7, 8], number of units per layer [512, 1024, 2048, 3072] and number of frames of acoustic data in the input l”
Machine Learning Paradigms for Speech Recognition: An Overview (2013, background): “First, the probabilistic segmental property of the target process is represented by the conditional probability [62]: , .”
Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups (2012, unknown): “…size, but even with this constraint it was impossible to train all possible combinations of number of hidden layers [1] [2] [3] [4] [5] [6] [7] [8], number of units per layer [512] [1] [024] [2] [048] [3] [072], and number of frames of acoustic data in the input layer [7] [11] [15] [17] [27] [37].”