End-to-end Training Approaches For Discriminative Segmental Models

Abstract

Recent work on discriminative segmental models has shown that they can achieve competitive speech recognition performance, using features based on deep neural frame classifiers. However, segmental models can be more challenging to train than standard frame-based approaches. While some segmental models have been successfully trained end to end, there is a lack of understanding of their training under different settings and with different losses. We investigate a model class based on recent successful approaches, consisting of a linear model that combines segmental features based on an LSTM frame classifier. Similarly to hybrid HMM-neural network models, segmental models of this class can be trained in two stages (frame classifier training followed by linear segmental model weight training), end to end (joint training of both frame classifier and linear weights), or with end-to-end fine-tuning after two-stage training. We study segmental models trained end to end with hinge loss, log

Stats

• citations5
• S2 citations—
• github stars0
• HF likes0
• heat score5.84
• arxiv keytang2016end

Related papers

• End-to-end Neural Segmental Models For Speech Recognition (2017)9.23
• Sequence Prediction With Neural Segmental Models (2017)0.00
• Segmental Recurrent Neural Networks For End-to-end Speech Recognition (2016)0.00
• Efficient Segmental Cascades For Speech Recognition (2016)0.00
• Sequence Segmentation Using Joint RNN And Structured Prediction Models (2016)7.81
• Multitask Learning With CTC And Segmental CRF For Speech Recognition (2017)0.00
• End-to-end Training Of A Neural HMM With Label And Transition Probabilities (2023)4.52
• Equivalence Of Segmental And Neural Transducer Modeling: A Proof Of Concept (2021)4.52