AudioFM paper on arXiv: recording-level contrast plus masked latent discrimination for general audio
iScienceLuvr · x · 2026-10-27
The arXiv page for AudioFM (arXiv:2610.34183, submitted 2026-10-25) by Yamin Li and 21 co-authors from Google Research/DeepMind, Vanderbilt, and University of Washington.
- Core idea: recording-level contrastive objectives only supervise pooled representations, not local time-frequency detail; AudioFM combines recording-level contrast with patch-wise masked latent discrimination — predicting masked patch representations from visible context and identifying matching full-view teacher targets among candidates from other recordings — plus weaker supervision from temporal neighbors.
- Findings: state-of-the-art across diverse audio domains in linear probing, fine-tuning, and zero-shot transfer, with stronger event detection, robust scene-level understanding, and strong domain generalizability.
Related event: DeepMind Releases AudioFM, Achieving SOTA on 10 of 12 Audio Tasks(2 posts)→
More from Research
- OpenAI claims 100+ longstanding math problems solved, 700+ results on GitHub in two months — haider1 · 2026-10-27
- Ben Lorica changes his mind on world models: if AI builds the robot, where's the value? — bigdata · 2026-10-27
- Study: just 10 minutes of AI assistance can erode motivation to solve problems alone — Polymarket · 2026-10-27
- Murea Labs' InWorld: AI-Generated World That Remembers Characters and Objects Turns Later — jiqizhixin · 2026-10-27
- GPT-6 moves a camera to insert a flower, yet burns an egg: reflections on the Astra report — liuziwei7 · 2026-10-27
- Eric Topol and colleagues in Nature: measuring immune system status in the clinic is about to change — EricTopol · 2026-10-27