2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information

2021 IEEE International Conference on Acoustics, Speech and Signal Processing

6-11 June 2021 • Toronto, Ontario, Canada

Extracting Knowledge from Information
Login Paper Search My Schedule Paper Index Help

My ICASSP 2021 Schedule

Note: Your custom schedule will not be saved unless you create a new account or login to an existing account.
  1. Create a login based on your email (takes less than one minute)
  2. Perform 'Paper Search'
  3. Select papers that you desire to save in your personalized schedule
  4. Click on 'My Schedule' to see the current list of selected papers
  5. Click on 'Printable Version' to create a separate window suitable for printing (the header and menu will appear, but will not actually print)

Paper Detail

Paper IDHLT-5.4
Paper Title TWO-STAGE TEXTUAL KNOWLEDGE DISTILLATION FOR END-TO-END SPOKEN LANGUAGE UNDERSTANDING
Authors Seongbin Kim, Inha University, South Korea; Gyuwan Kim, Seongjin Shin, NAVER Corp., South Korea; Sangmin Lee, Inha University, South Korea
SessionHLT-5: Language Understanding 1: End-to-end Speech Understanding 1
LocationGather.Town
Session Time:Wednesday, 09 June, 13:00 - 13:45
Presentation Time:Wednesday, 09 June, 13:00 - 13:45
Presentation Poster
Topic Human Language Technology: [HLT-UNDE] Spoken Language Understanding and Computational Semantics
IEEE Xplore Open Preview  Click here to view in IEEE Xplore
Abstract End-to-end approaches open a new way for more accurate and efficient spoken language understanding (SLU) systems by alleviating the drawbacks of traditional pipeline systems. Previous works exploit textual information for an SLU model via pre-training with automatic speech recognition or fine-tuning with knowledge distillation. To utilize textual information more effectively, this work proposes a two-stage textual knowledge distillation method that matches utterance-level representations and predicted logits of two modalities during pre-training and fine-tuning, sequentially. We use vq-wav2vec BERT as a speech encoder because it captures general and rich features. Furthermore, we improve the performance, especially in a low-resource scenario, with data augmentation methods by randomly masking spans of discrete audio tokens and contextualized hidden representations. Consequently, we push the state-of-the-art on the Fluent Speech Commands, achieving 99.7% test accuracy in full dataset setting and 99.5% in 10% subset setting. Throughout the ablation studies, we empirically verify that all used methods are crucial to the final performance, providing the best practice for spoken language understanding. Code to reproduce our results will be available upon publication.