A generalized hidden Markov model with state-conditioned trend functions of time for the speech signal

By Li Deng. 1992

How researchers see this paper

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

1. Li Deng
2. Heiga Zen
3. Dong Yu
4. Keiichi Tokuda
5. Junichi Yamagishi
6. Alan W. Black
7. Xu Tan
8. Zhizheng Wu

Citations in the corpus

• Deep Learning: Methods and Applications (2014, background): “Specifically, the early work described in [71, 72, 83, 84, 99, 274] generalized and extended the conventional shallow and conditionally independent HMM structure by imposing dynamic constraints, in the form of polynomial trajectory, on the HMM parameters.”
• A tutorial survey of architectures, algorithms, and applications for deep learning (2014, neutral): “Specifically, the early work described in [54–59] generalized and extended the conventional shallow and li deng conditionally independent HMM structure by imposing dynamic constraints, in the form of polynomial trajectory, on the HMM parameters.”
• Statistical Parametric Speech Synthesis (2007, methodology): “To alleviate the problem with piece-wise constant statistics, Dines and Sridharan applied trended HMMs (Deng, 1992), which included linearly time-varying functions in their state-output probabilities, to statistical parametric synthesis (Dines and Sridharan, 2001).”
• Structured speech modeling (2006, background): “• Polynomial trended HMM: — Observable trend functions: This is the simplest trended HMM where there is no uncertainty in polynomial coefficients [18], [21], [22], [27], [35].”
• An Overview of Deep-Structured Learning for Information Processing (2011, unknown): “Specifically, the early work described in (Deng 1992, 1993; Deng et al., 1994; Ostendorf et al., 1996, Deng and Sameti, 1996) generalized and extended the conventional shallow HMM structure by imposing dynamic constraints, in the form of polynomial trajectory, on the HMM parameters.”
• Computational Models for Speech Production (2018, background): “[3, 14, 11]); 2) multi-region dynamic system model whose sample paths are piecewise, recursively defined stochastic trajectories (e.”
• A bidirectional target-filtering model of speech coarticulation and reduction: two-stage implementation for phonetic recognition (2006, unknown): “Earlier work directly represented speech dynamics in the observed acoustic domain [5], [16], [19], [24].”
• Dynamic Speech Models: Theory, Algorithms, and Applications (2006, background): “However, a number of approaches, notably known as the stochastic segment model, segmental HMM, and trended trajectory models [14,40,41], which are intended to specifically overcome some of these inconsistent assumptions, have not delivered significant and consistent performance improvements, especia”
• Deep segmental neural networks for speech recognition (2013, methodology): “It has been known for decades that modeling speech using the conventional HMM has several limitations as analyzed in [9, 10, 11].”
• Machine Learning Paradigms for Speech Recognition: An Overview (2013, background): “This class of beyond-HMM models have been variably called stochastic segment model [31], [32], trended or nonstationary-state HMM [33], [34], trajectory segmental model [32], [35], trajectory HMMs [36], [37], stochastic trajectory models [38], hidden dynamic models [39]–[45], buried Markov models…”
• Speech trajectory discrimination using the minimum classification error learning (1998, methodology): “Abstract—In this paper, we extend the maximum likelihood (ML) training algorithm to the minimum classification error (MCE) training algorithm for discriminatively estimating the state-dependent polynomial coefficients in the stochastic trajectory model or the trended hidden Markov model (HMM) origin”