Google DeepMind has developed Perch 2.0, a bioacoustics foundation model initially trained on terrestrial animal vocalizations, which has demonstrated surprising effectiveness in underwater acoustic analysis. This model, detailed in a paper presented at NeurIPS 2025, can be utilized for transfer learning to classify whale vocalizations and identify other marine species. Researchers can leverage Perch 2.0 with tools like Google Colab and Google Cloud to create custom classifiers for underwater sound analysis, significantly reducing the computational resources and time required compared to building models from scratch. AI
IMPACT Enables more efficient and scalable insights into marine ecosystems by leveraging terrestrial animal sound models for underwater acoustic analysis.
RANK_REASON The item describes a new foundation model for bioacoustics and its application in a research paper, fitting the research bucket. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Google AI / Research →
- Google Cloud
- Google Colab
- Google DeepMind
- Google Research
- humpback whale
- National Oceanic and Atmospheric Administration
- Neurips 2025
- NOAA NCEI Passive Acoustic Data Archive
- Perch 2.0
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
