Adding Connectionist Temporal Summarization into Conformer to Improve Its Decoder Efficiency For Speech Recognition

TL;DR

This paper introduces a connectionist temporal summarization method integrated into Conformer models to enhance speech recognition decoding efficiency, reducing computational load without sacrificing accuracy.

Contribution

The paper proposes a novel CTS technique for Conformer that decreases decoding operations and improves efficiency while maintaining or improving ASR accuracy.

Findings

Decoding budget reduced by up to 20% on LibriSpeech

Decoding budget reduced by 11% on FluentSpeech

WER reduced by 6% at beam width 1

Abstract

The Conformer model is an excellent architecture for speech recognition modeling that effectively utilizes the hybrid losses of connectionist temporal classification (CTC) and attention to train model parameters. To improve the decoding efficiency of Conformer, we propose a novel connectionist temporal summarization (CTS) method that reduces the number of frames required for the attention decoder fed from the acoustic sequences generated by the encoder, thus reducing operations. However, to achieve such decoding improvements, we must fine-tune model parameters, as cross-attention observations are changed and thus require corresponding refinements. Our final experiments show that, with a beamwidth of 4, the LibriSpeech's decoding budget can be reduced by up to 20% and for FluentSpeech data it can be reduced by 11%, without losing ASR accuracy. An improvement in accuracy is even found for the LibriSpeech "test-other" set. The word error rate (WER) is reduced by 6\% relative at the beam width of 1 and by 3% relative at the beam width of 4.