Software & Data Downloads — NABEATs

Noise-Aware Audio Representation Learning for extracting clean self-supervised learning (SSL) representations from noisy input by using auxiliary noise information.

PyTorch code for extracting clean self-supervised learning (SSL) representations from noisy input by using auxiliary noise information. This repository contains the pre-trained models and training recipes used in our IWAENC 2026 and DCASE 2026 papers.

    •  Fujimura, T., Masuyama, Y., Wichern, G., Boeddeker, C., Richter, J., Le Roux, J., "NABEATs: Noise-Aware Audio Representation Learning", International Workshop on Acoustic Signal Enhancement (IWAENC), September 2026.
      BibTeX TR2026-124 PDF Software
      • @inproceedings{Fujimura2026sep,
      • author = {Fujimura, Takuya and Masuyama, Yoshiki and Wichern, Gordon and Boeddeker, Christoph and Richter, Julius and {Le Roux}, Jonathan},
      • title = {{NABEATs: Noise-Aware Audio Representation Learning}},
      • booktitle = {International Workshop on Acoustic Signal Enhancement (IWAENC)},
      • year = 2026,
      • month = sep,
      • url = {https://www.merl.com/publications/TR2026-124}
      • }

    Access software at https://github.com/merlresearch/noise-aware-audio-ssl.