Topics

The special session topics are around the theme of building speech recognition systems on large pre-trained model, on languages which are underrepresented udring pre-training stage. The following technical topics using large pretrained models are allowed for submission as part of this special session

• Novel fine tuning techniques

• Adapters and prompt tuning

• re-scoring approaches 

• data augmentation

• zero-shot/few-shot transfer

• test-time adaptation

• pseudo-label training

• ensembling

Further, we also allow submissions around analysis and datasets 

• Detailed analysis of performance with pretrained models with under-represented datasets

• Open sourcing new low-resource datasets and studying the performance of large pre-trained models

The topics listed are not exhaustive, we encourage authors to submit on related areas as well. We recommend getting in touch with the organisers to ensure that the submission is within the theme of the special session.