Hidden conditional random fields for phone classification

By Asela Gunawardana, Milind Mahajan, A. Acero, John C. Platt. Interspeech, 2005

366 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/ad4cfdb0c0bb62bf1fa99cb3d0519cadd9cbc160

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

1. Li Deng
2. Dong Yu
3. Yoshua Bengio
4. Jianfeng Gao
5. Percy Liang
6. Mohit Bansal
7. Noah A. Smith
8. Jinyu Li

In Signal Processing

1. Li Deng
2. Dong Yu
3. Jinyu Li
4. Shinji Watanabe
5. S. Renals
6. George E. Dahl
7. Xiao Li
8. Yifan Gong

Citations in the corpus

• Machine Learning Paradigms for Speech Recognition: An Overview (2013, methodology): “For example, when applying (26) to Markov random fields, we obtain the training objective of conditional random fields (CRFs) [70]: (27) The partition function is a normalization factor. is a weight vector and is a vector of feature functions referred to as a feature vector.”
• Latent-Dynamic Discriminative Models for Continuous Gesture Recognition (2007, methodology): “In the speech community, a similar model was applied to phone classification [7].”
• Discriminative Input Stream Combination for Conditional Random Field Phone Recognition (2009, methodology): “, MFCC-based Gaussian sufficient statistics [10] or relative distance to Gaussian means [11]) that we intend to explore in future work.”
• Optimization Algorithms and Applications for Speech and Language Processing (2013, methodology): “Thus, by applying the Taylor series argument mentioned above, we can show that for all sufficiently large—we are guaranteed to see ascent in .”
• SCARF: A Segmental CRF Speech Recognition System (2009, background): “Hidden CRFs (HCRFs) [5] address this issue by summing over all labelings consistent with a known or hypothesized word sequence.”
• A segmental CRF approach to large vocabulary continuous speech recognition (2009, positive): “SEGMENTAL CRF FORMULATION Segmental Conditional Random Fields -also known as Semi-Markov Random Fields [16] or SCRFs -underlie our approach.”
• Context-Dependent Pre-Trained Deep Neural Networks for Large-Vocabulary Speech Recognition (2012, methodology): “The recently proposed CRF [18]–[20] and HCRF [21], [22] models use log-linear models to replace GMM-HMMs.”
• Segmental Recurrent Neural Networks for End-to-End Speech Recognition (2016, background): “, [2, 3, 4, 5]; however these approaches have not yet improved speech recognition accuracy over HMMs.”
• Exploiting deep neural networks for detection-based speech recognition (2013, background): “Finally, (hidden) conditional random fields have also been proposed [17,57].”
• An Introduction to Conditional Random Fields (2010, background): “Since then, linear-chain CRFs have been applied to many problems in natural language processing, including named-entity recognition [72], feature induction for NER [71], shallow parsing [108, 120], identifying protein names in biology abstracts [107], segmenting addresses in Web pages [26], informat”
• Discriminative learning in sequential pattern recognition (2008, background): “IEEE SIGNAL PROCESSING MAGAZINE [21] SEPTEMBER 2008 In (25), X 1 . . .”
• Discriminative Training for Large-Vocabulary Speech Recognition Using Minimum Classification Error (2007, background): “0% obtained for Quickprop is the best result known to us for a standard HMM on this well-known task [23], [60]–[62].”
• Deep-structured hidden conditional random fields for phonetic recognition (2010, methodology): “Parameter estimation at the final layer can thus be carried out in a supervised manner in the same way as explained in [5][6][7][9].”
• Robust Support Vector Machine Training via Convex Outlier Ablation (2006, methodology): “Interestingly, the original proposal for soft margin SVMs (Cortes & Vapnik 1995) considered alternative losses based on the transformation loss ( w , x , y ) = hinge ( w , x , y ) p .”
• Investigations on an EM-Style Optimization Algorithm for Discriminative Training of HMMs (2013, background): “2 illustrates this inequality for two points .”
• Hidden Conditional Random Fields for Gesture Recognition (2006, methodology): “Our hidden state conditional random field (HCRF) model can be used either as a gesture class detector, where a single class is discriminatively trained against all other gestures, or as a multi-way gesture classifier, where discriminative models for multiple gestures are simultaneously trained.”
• Equivalence of Generative and Log-Linear Models (2011, methodology): “The normalization constant is computed by running over all tag strings of length .”