Self-supervised Contrastive Cross-Modality Representation Learning for Spoken Question Answering

Abstract

Spoken question answering (SQA) requires fine-grained understanding of both spoken documents and questions for the optimal answer prediction. In this paper, we propose novel training schemes for spoken question answering with a self-supervised training stage and a contrastive representation learning stage. In the self-supervised stage, we propose three auxiliary self-supervised tasks, including utterance restoration, utterance insertion, and question discrimination, and jointly train the model to capture consistency and coherence among speech documents without any additional data or annotations. We then propose to learn noise-invariant utterance representations in a contrastive objective by adopting multiple augmentation strategies, including span deletion and span substitution. Besides, we design a Temporal-Alignment attention to semantically align the speech-text clues in the learned common space and benefit the SQA tasks. By this means, the training schemes can more effectively guide the generation model to predict more proper answers. Experimental results show that our model achieves state-of-the-art results on three SQA benchmarks.

References in corpus (11)

• Improved Baselines with Momentum Contrastive Learning
• Contrastive Multiview Coding
• Mockingjay: Unsupervised Speech Representation Learning with Deep Bidirectional Transformer Encoders
• Structure-sensitive Multi-scale Deep Neural Network for Low-Dose CT Denoising
• CLEAR: Contrastive Learning for Sentence Representation
• Data Augmenting Contrastive Learning of Speech Representations in the Time Domain
• FlowQA: Grasping Flow in History for Conversational Machine Comprehension
• Spoken SQuAD: A Study of Mitigating the Impact of Speech Recognition Errors on Listening Comprehension
• Towards Data Distillation for End-to-end Spoken Conversational Question Answering
• CSTNet: Contrastive Speech Translation Network for Self-Supervised Speech Representation Learning
• Self-supervised pre-training and contrastive representation learning for multiple-choice video QA