{"id":"arxiv/2603.27414","url":"https://arxiv.org/abs/2603.27414","tags":["math.ST","cs.AI"],"score":0,"title":"Multiple-Prediction-Powered Inference","domain":"math.ST","source":"arxiv","authors":["Charlie Cowen-Breen","Alekh Agarwal","Stephen Bates","William W. Cohen","Jacob Eisenstein","Amir Globerson","Adam Fisch"],"reviews":0,"subject":"math.ST","summary":"Statistical estimation often involves tradeoffs between expensive, high-quality measurements and a variety of lower-quality proxies. We introduce Multiple-Prediction-Powered Inference (MultiPPI): a general framework for constructing statistically efficient estimates by optimally allocating resources across these diverse data sources. This work provides theoretical guarantees about the minimax optimality, finite-sample performance, and asymptotic normality of the MultiPPI estimator. Through experiments across three diverse large language model (LLM) evaluation scenarios, we show that MultiPPI…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-28","reactions":{}}