Contrastive Learning for Improving ASR Robustness in Spoken Language Understanding
This work addresses robustness in SLU for better human-machine interactions, but it is incremental as it builds on existing contrastive learning methods.
The paper tackles the problem of ASR errors degrading spoken language understanding performance by learning robust utterance representations using contrastive learning, achieving improved results on three benchmark datasets.
Spoken language understanding (SLU) is an essential task for machines to understand human speech for better interactions. However, errors from the automatic speech recognizer (ASR) usually hurt the understanding performance. In reality, ASR systems may not be easy to adjust for the target scenarios. Therefore, this paper focuses on learning utterance representations that are robust to ASR errors using a contrastive objective, and further strengthens the generalization ability by combining supervised contrastive learning and self-distillation in model fine-tuning. Experiments on three benchmark datasets demonstrate the effectiveness of our proposed approach.