Model merging boosts bioacoustic foundation model NatureLM-audio F1 by 200%+
emaros96 · x · 2026-10-09
Researcher Yannis Pierri discusses work with the Earth Species team on NatureLM-audio, a bioacoustic foundation model. Using model merging, they improved the model's instruction-following while retaining its expertise in animal sounds, achieving over 200% relative F1 improvement on unseen taxonomic families. The interview covers the method and the broader questions driving ML research on animal communication.
More from Research
- Domingos Cites New Result Claiming RL Is 'Basically Useless' vs Supervised Learning — pmddomingos · 2026-10-23
- Wolfram Opens Rule 30 Prizes to AIs: $30,000 for Three Open Problems — cephaloform · 2026-10-23
- Stanford and Databricks use AI agents to reproduce medical findings and test analytical robustness — ChengleiSi · 2026-10-23
- EMNLP paper: diffusion LMs internally encode latent denoising timestep representations — gsarti_ · 2026-10-23
- KU Leuven and VIB open two Group Leader professorships in AI for Biology, start 2027 — jpillowtime · 2026-10-23
- Subspace Tuning curbs LLM catastrophic forgetting by freezing principal singular directions — scottniekum · 2026-10-23