Speech Recognition with Deep Recurrent Neural Networks

Abstract

Recurrent neural networks (RNNs) are a powerful model for sequential data. End-to-end training methods such as Connectionist Temporal Classification make it possible to train RNNs for sequence labelling problems where the input-output alignment is unknown. The combination of these methods with the Long Short-term Memory RNN architecture has proved particularly fruitful, delivering state-of-the-art results in cursive handwriting recognition. However RNN performance in speech recognition has so far been disappointing, with better results returned by deep feedforward networks. This paper investigates \emph{deep recurrent neural networks}, which combine the multiple levels of representation that have proved so effective in deep networks with the flexible use of long range context that empowers RNNs. When trained end-to-end with suitable regularisation, we find that deep Long Short-term Memory RNNs achieve a test set error of 17.7% on the TIMIT phoneme recognition benchmark, which to our knowledge is the best recorded score.

To appear in ICASSP 2013

References in corpus (1)

• Sequence Transduction with Recurrent Neural Networks

Cited by in corpus (10)

• Learning Stochastic Recurrent Networks
• On Fast Dropout and its Applicability to Recurrent Networks
• Detrending Exoplanetary Transit Light Curves with Long Short-Term Memory Networks
• Bidirectional Recurrent Neural Networks as Generative Models - Reconstructing Gaps in Time Series
• Lie Access Neural Turing Machine
• Deep Transfer Learning for Classification of Variable Sources
• Distributed Training Large-Scale Deep Architectures
• Omni SCADA Intrusion Detection Using Deep Learning Algorithms
• Development of Automatic Speech Recognition for Kazakh Language using Transfer Learning
• Variational inference of latent state sequences using Recurrent Networks