Nature Communications

Sufficient is better than optimal for training neural networks

• Irina Babayan
• Hazhir Aliahmadi
• Greg van Anders

2025-12-04

Full text

DOI https://doi.org/10.1038/s41467-025-66983-3