Linear predictive hidden Markov models and the speech signal

By A. Poritz. 1982

Subfields: Natural Language Processing, Speech & Audio

How researchers see this paper

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

In Natural Language Processing

In Speech & Audio

Citations in the corpus

WaveNet: A Generative Model for Raw Audio (2016, background): “• Gaussian process assumption; The conventional generative models are based on Gaussian process (Itakura & Saito, 1970; Imai & Furuichi, 1988; Poritz, 1982; Juang & Rabiner, 1985; Kameoka et al., 2010; Tokuda & Zen, 2015; 2016).”
Estimating phoneme class conditional probabilities from raw speech signal using convolutional neural networks (2013, background): “In one of the earliest work, Poritz proposed an approach where the speech signal is modeled by a linear prediction HMM [5].”
Acoustic modeling with deep neural networks using raw time signal for LVCSR (2014, unknown): “Some works on processing of raw speech signal make use of other models such as linear predictive models [11][12] or SVMs [13].”
Autoregressive Models for Statistical Parametric Speech Synthesis (2013, methodology): “This is distinct from the hidden filter HMM [9], [10] for which the observations are waveform samples.”
Hidden Markov processes (2002, methodology): “Parametrization of each covariance matrix as a matrix of an autoregressive process of order [149] was studied in [254], [255], [175].”
MOSAIC Model for Sensorimotor Learning and Control (2001, methodology): “Since the maximization step for linear parameters is a maximum likelihood estimation weighted by γt(i) (Poritz, 1982; Fraser & Dimitriadis, 1993), new estimates of W and σ (Ŵ and σ̂ , respectively) are computed as in equations 3.4 and 3.5, Ŵi = (Y′γ (i)Y)−1Y′X (3.4) σ̂ 2i = (X − YŴi)′γ (i)(X −…”
Learning grounded finite-state representations from unstructured demonstrations (2015, background): “Autoregressive HMMs [Poritz, 1982] have been proposed to represent these dependencies by adding causal links between observations in the HMM, but it can be difficult to represent the transition structure when the observation space is large or continuous.”
Automatic Speech Recognition - A Brief History of the Technology Development (2004, background): “Baum’s doubly stochastic process started to find applications in the speech area, initially in speaker identification systems, in the late 1970’s [40-41].”
A Systematic Review of Hidden Markov Models and Their Applications (2020, background): “Poritz [12] proposed a linear predictive HMM for analyzing the speech signals.”
A tutorial on hidden Markov models and selected applications in speech recognition (1989, background): “This is the class of auto-regressive HMMs [27, [28].”