End-to-End Neural Segmental Models for Speech Recognition
arXiv:1708.00531 · doi:10.1109/JSTSP.2017.2752462
Abstract
Segmental models are an alternative to frame-based models for sequence prediction, where hypothesized path weights are based on entire segment scores rather than a single frame at a time. Neural segmental models are segmental models that use neural network-based weight functions. Neural segmental models have achieved competitive results for speech recognition, and their end-to-end training has been explored in several studies. In this work, we review neural segmental models, which can be viewed as consisting of a neural network-based acoustic encoder and a finite-state transducer decoder. We study end-to-end segmental models with different weight functions, including ones based on frame-level neural classifiers and on segmental recurrent neural networks. We study how reducing the search space size impacts performance under different weight functions. We also compare several loss functions for end-to-end training. Finally, we explore training approaches, including multi-stage vs. end-to-end training and multitask training that combines segmental and frame-level losses.
References in corpus (6)
- Recurrent Neural Network Regularization
- Attention-Based Models for Speech Recognition
- Sequence Transduction with Recurrent Neural Networks
- Wav2Letter: an End-to-End ConvNet-based Speech Recognition System
- Segmental Recurrent Neural Networks
- Multitask Learning with CTC and Segmental CRF for Speech Recognition
Cited by in corpus (9)
- Letter-Based Speech Recognition with Gated ConvNets
- Time-Contrastive Learning Based Deep Bottleneck Features for Text-Dependent Speaker Verification
- Frame-level speaker embeddings for text-independent speaker recognition and analysis of end-to-end model
- ContextNet: A Click-Through Rate Prediction Framework Using Contextual information to Refine Feature Embedding
- On Bottleneck Features for Text-Dependent Speaker Verification Using X-vectors
- From Semi-supervised to Almost-unsupervised Speech Recognition with Very-low Resource by Jointly Learning Phonetic Structures from Audio and Text Embeddings
- On The Inductive Bias of Words in Acoustics-to-Word Models
- Whole-Word Segmental Speech Recognition with Acoustic Word Embeddings
- Phonetic-and-Semantic Embedding of Spoken Words with Applications in Spoken Content Retrieval