5 papers
Multiple-Prediction-Powered Inference
Charlie Cowen-Breen, Alekh Agarwal, Stephen Bates +4
Statistical estimation often involves tradeoffs between expensive, high-quality measurements and a variety of lower-quality proxies. We introduce Multiple-Prediction-Powered Infere…
An Annotated Dataset of Errors in Premodern Greek and Baselines for Detecting Them
Creston Brooks, Johannes Haubold, Charlie Cowen-Breen +5
As premodern texts are passed down over centuries, errors inevitably accrue. These errors can be challenging to identify, as some have survived undetected for so long precisely bec…
Towards Probabilistically-Sound Beam Search with Masked Language Models
Creston Brooks, Robert Calef, Charlie Cowen-Breen +1
Beam search with masked language models (MLMs) is challenging in part because joint probability distributions over sequences are not readily available, unlike for autoregressive mo…
Logion: Machine Learning for Greek Philology
Charlie Cowen-Breen, Creston Brooks, Johannes Haubold +1
This paper presents machine-learning methods to address various problems in Greek philology. After training a BERT model on the largest premodern Greek dataset used for this purpos…
Pattern Problems related to the Arithmetic Kakeya Conjecture
Charlie Cowen-Breen, Elene Karangozishvili, Narmada Varadarajan +1
We study a variety of problems about homothets of sets related to the Kakeya conjecture. In particular, we show many of these problems are equivalent to the arithmetic Kakeya conje…