First-Pass Large Vocabulary Continuous Speech Recognition using Bi-Directional Recurrent DNNs

Abstract

We present a method to perform first-pass large vocabulary continuous speech recognition using only a neural network and language model. Deep neural network acoustic models are now commonplace in HMM-based speech recognition systems, but building such systems is a complex, domain-specific task. Recent work demonstrated the feasibility of discarding the HMM sequence modeling framework by directly predicting transcript text from audio. This paper extends this approach in two ways. First, we demonstrate that a straightforward recurrent neural network architecture can achieve a high level of accuracy. Second, we propose and evaluate a modified prefix-search decoding algorithm. This approach to decoding enables first-pass speech recognition with a language model, completely unaided by the cumbersome infrastructure of HMM-based systems. Experiments on the Wall Street Journal corpus demonstrate fairly competitive word error rates, and the importance of bi-directional network recurrence.

Cited by in corpus (16)

• cuDNN: Efficient Primitives for Deep Learning
• An Actor-Critic Algorithm for Sequence Prediction
• Self-Attention Networks for Connectionist Temporal Classification in Speech Recognition
• Exploring Neural Transducers for End-to-End Speech Recognition
• Radio2Text: Streaming Speech Recognition Using mmWave Radio Signals
• Emphasizing Unseen Words: New Vocabulary Acquisition for End-to-End Speech Recognition
• Attention-based Wav2Text with Feature Transfer Learning
• Improving End-to-End Speech Recognition with Policy Learning
• End-to-end attention-based distant speech recognition with Highway LSTM
• Back from the future: bidirectional CTC decoding using future information in speech recognition
• KT-Speech-Crawler: Automatic Dataset Construction for Speech Recognition from YouTube Videos
• An Investigation of Enhancing CTC Model for Triggered Attention-based Streaming ASR
• Phoneme Level Language Models for Sequence Based Low Resource ASR
• Topic Model Robustness to Automatic Speech Recognition Errors in Podcast Transcripts
• Exploiting Nontrivial Connectivity for Automatic Speech Recognition
• Train your classifier first: Cascade Neural Networks Training from upper layers to lower layers