Google's Gemini model has been adapted to classify astronomical events like supernovae with high accuracy using a few-shot learning approach. By providing Gemini with only 15 annotated examples and concise instructions, researchers achieved 93% accuracy across three astronomical surveys, matching the performance of specialized models that require vast datasets. Crucially, Gemini can also provide plain-language explanations for its classifications and assign an interest score, significantly aiding astronomers in sifting through the massive data generated by modern telescopes. AI
IMPACT Enhances AI's utility in scientific research by enabling complex data analysis with minimal training data, potentially accelerating discovery.
RANK_REASON Publication of a research paper detailing a new application and methodology for a large language model. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Google AI / Research →
- Dr. Fiorenzo Stoppa
- Gemini
- Google Cloud
- MeerLICHT
- Nature Astronomy
- Pan-STARRS
- Turan Bulmus
- University of Oxford
- Vera C. Rubin Observatory
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
