Speech recognition with deep recurrent neural networks

TLDR

Recurrent neural networks, especially when trained end‑to‑end with Connectionist Temporal Classification and Long Short‑Term Memory units, have achieved state‑of‑the‑art results in tasks such as cursive handwriting recognition, yet their performance in speech recognition has historically lagged behind deep feedforward networks. This study explores deep recurrent neural networks that merge multi‑layer representation learning with long‑range contextual modeling. The authors train deep Long Short‑Term Memory RNNs end‑to‑end, applying regularisation techniques to prevent overfitting. The resulting deep LSTM RNNs achieve a 17.7% error rate on the TIMIT phoneme recognition benchmark, the best score reported to date.

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 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.