Automatic prosody prediction and detection with Conditional Random Field (CRF) models

By Yao Qian, Zhizheng Wu, Xuezhe Ma, Frank K. Soong. 2010

54 citations recorded by the source corpus.

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

Citations in the corpus

A Survey on Neural Speech Synthesis (2021, methodology): “TTS Text Analysis Char→Linguistic TN [310, 223, 430], G2P [403, 320, 321] Prosody Prediction [312, 137, 277, 251] Unified Model [252, 437] DeepVoice 1/2 [8, 84]”
Unified Mandarin TTS Front-end Based on Distilled BERT Model (2020, methodology): “Systems for prosodic structure prediction: CRF: We implement CRF model with open source tool CRF++1.”
DurIAN: Duration Informed Attention Network for Speech Synthesis (2020, background): “Therefore, it is still more desirable to have a separate model for learning word and prosody boundaries from larger dataset [13, 14, 15] in tasks such as Mandarin speech synthesis.”
Automatic prosody prediction for Chinese speech synthesis using BLSTM-RNN and embedding features (2015, background): “To our knowledge, the best reported results were achieved with CRF due to its ability of relaxing strong model independence assumption and solving the label bias problem [1, 10].”
Emotion-Aware Prosodic Phrasing for Expressive Text-to-Speech (2023, background): “For the second category, researchers tried to build the prosodic phrasing model with conditional random fields (CRF) [12], deep neural networks (DNNs) [13], recurrent neural networks (RNNs) [14], bidirectional long short-term memory (BiLSTM) [15], and self-attention-based transformer network [16], t”
RobuTrans: A Robust Transformer-Based Text-to-Speech Model (2020, methodology): “The prosodic feature includes tone and break index, which are predicted by a conditional random ﬁeld (CRF) model with syntactic and contextual information as in Qian et al. (2010).”
A Unified Sequence-to-Sequence Front-End Model for Mandarin Text-to-Speech Synthesis (2019, methodology): “Typical prosody prediction methods include rule-based models and statistical models like CRF [7], and RNN [5].”
Review of end-to-end speech synthesis technology based on deep learning (2021, methodology): “For example, Tahon et al. [203] added a pronunciation adaptive framework based on CRF between text front-end and TTS model to generate different styles of speech.”
Self-attention Based Prosodic Boundary Prediction for Chinese Speech Synthesis (2019, methodology): “BLSTM-CRF: BLSTM-CRF based model which uses word embeddings as input and achieves state-of-the-art results in [11] is used as our baseline.”