Deep ensembles
Deep ensembles are a machine learning technique that combines multiple independently trained neural networks to improve predictive performance and uncertainty estimation. They are widely used to enhance model robustness and provide more reliable confidence measures in predictions.