Do We Still Need Automatic Speech Recognition for Spoken Language Understanding?

TL;DR

This paper evaluates whether recent speech representation learning can replace traditional ASR in spoken language understanding tasks, finding that learned features outperform ASR transcripts in classification but not in translation.

Contribution

It demonstrates the potential of wav2vec 2.0 features to replace ASR in certain SLU tasks, highlighting their robustness and effectiveness.

Findings

Learned speech features outperform ASR transcripts in classification tasks.

ASR transcripts are still better for machine translation.

Wav2vec 2.0 representations are robust to out-of-vocabulary words.

Abstract

Spoken language understanding (SLU) tasks are usually solved by first transcribing an utterance with automatic speech recognition (ASR) and then feeding the output to a text-based model. Recent advances in self-supervised representation learning for speech data have focused on improving the ASR component. We investigate whether representation learning for speech has matured enough to replace ASR in SLU. We compare learned speech features from wav2vec 2.0, state-of-the-art ASR transcripts, and the ground truth text as input for a novel speech-based named entity recognition task, a cardiac arrest detection task on real-world emergency calls and two existing SLU benchmarks. We show that learned speech features are superior to ASR transcripts on three classification tasks. For machine translation, ASR transcripts are still the better choice. We highlight the intrinsic robustness of wav2vec 2.0 representations to out-of-vocabulary words as key to better performance.