Calibrated ensembles can mitigate accuracy tradeoffs under distribution shift.

Ananya Kumar Tengyu Ma Percy Liang Aditi Raghunathan

Published in: CoRR (2022)

Keyphrases

high accuracy
computational cost
prediction accuracy
active learning
highly accurate
ensemble methods
decision trees
probability distribution
power law
neural network
correlation coefficient
precision and recall
random variables
feature space
computational complexity
computer vision
search engine
genetic algorithm