SADA: Safe and Adaptive Aggregation of Multiple Black-Box Predictions in Semi-Supervised Learning
Researchers have developed SADA, a new method for safely and adaptively aggregating predictions from multiple black-box models in semi-supervised learning scenarios. This approach guarantees performance no worse than using labeled data alone and can achieve optimal efficiency if any single prediction is perfect. The method has been demonstrated through simulations and real-world data analyses, with an accompanying R package available for implementation. AI
IMPACT Enhances semi-supervised learning by enabling more robust aggregation of diverse model predictions.