Self-supervised learning has been proved to benefit a wide range of speech processing tasks, such as speech recognition/translation, speaker verification and diarization, etc. However, most of current approaches are computationally expensive. In this paper, we propose a simplified and more efficient self-supervised learning framework termed as NeMo Encoder for Speech Tasks (NEST). Specifically, we adopt the FastConformer architecture with 8x sub-sampling rate, which is faster than Transformer or Conformer architectures. Instead of clustering-based quantization, we use fixed random projection for its simplicity and effectiveness. We also implement a generalized noisy speech augmentation that teaches the model to disentangle the main speaker from noise or other speakers. Experiments show that \model improves over existing self-supervised models and achieves new state-of-the-art performance on a variety of speech processing tasks, such as speech recognition/translation, speaker diarization, spoken language understanding, etc. Code and checkpoints are publicly available via NVIDIA NeMo framework.
@article{arxiv.2408.13106,
title = {NEST: Self-supervised Fast Conformer as All-purpose Seasoning to Speech Processing Tasks},
author = {He Huang and Taejin Park and Kunal Dhawan and Ivan Medennikov and Krishna C. Puvvada and Nithin Rao Koluguri and Weiqing Wang and Jagadeesh Balam and Boris Ginsburg},
journal= {arXiv preprint arXiv:2408.13106},
year = {2025}
}