DeepMind Releases AudioFM, Achieving SOTA on 10 of 12 Audio Tasks

Google DeepMind has released AudioFM, a self-supervised framework for universal audio representation combining recording-level contrastive learning and masked discrimination. It achieves best results on 10 of 12 audio tasks, surpassing prior SSL methods in medical auscultation.

2026-10-27 ~ 2026-10-27 · 2 related posts

• DeepMind's AudioFM hits best results on 10 of 12 audio tasks, beats prior SSL on heart murmur by +4.7 F1 — iScienceLuvr · 2026-10-27
• AudioFM paper on arXiv: recording-level contrast plus masked latent discrimination for general audio — iScienceLuvr · 2026-10-27