Prediction-powered inference

AN Angelopoulos, S Bates, C Fannjiang, MI Jordan… - Science, 2023‏ - science.org
Prediction-powered inference is a framework for performing valid statistical inference when
an experimental dataset is supplemented with predictions from a machine-learning system …

Adaptive and robust multi-task learning

Y Duan, K Wang - The Annals of Statistics, 2023‏ - projecteuclid.org
Adaptive and robust multi-task learning Page 1 The Annals of Statistics 2023, Vol. 51, No. 5,
2015–2039 https://doi.org/10.1214/23-AOS2319 © Institute of Mathematical Statistics, 2023 …

Learning from similar linear representations: Adaptivity, minimaxity, and robustness

Y Tian, Y Gu, Y Feng - arxiv preprint arxiv:2303.17765, 2023‏ - arxiv.org
Representation multi-task learning (MTL) has achieved tremendous success in practice.
However, the theoretical understanding of these methods is still lacking. Most existing …

Doubly high-dimensional contextual bandits: An interpretable model for joint assortment-pricing

J Cai, R Chen, MJ Wainwright, L Zhao - arxiv preprint arxiv:2309.08634, 2023‏ - arxiv.org
Key challenges in running a retail business include how to select products to present to
consumers (the assortment problem), and how to price products (the pricing problem) to …

COMMUTE: communication-efficient transfer learning for multi-site risk prediction

T Gu, PH Lee, R Duan - Journal of biomedical informatics, 2023‏ - Elsevier
Objectives: We propose a communication-efficient transfer learning approach (COMMUTE)
that effectively incorporates multi-site healthcare data for training a risk prediction model in a …

Speed up the cold-start learning in two-sided bandits with many arms

M Bayati, J Cao, W Chen - arxiv preprint arxiv:2210.00340, 2022‏ - arxiv.org
Multi-armed bandit (MAB) algorithms are efficient approaches to reduce the opportunity cost
of online experimentation and are used by companies to find the best product from …

Estimation and inference for transfer learning with high-dimensional quantile regression

J Huang, M Wang, Y Wu - arxiv preprint arxiv:2211.14578, 2022‏ - arxiv.org
Transfer learning has become an essential technique to exploit information from the source
domain to boost performance of the target task. Despite the prevalence in high-dimensional …

Transportability for bandits with data from different environments

A Bellot, A Malek, S Chiappa - Advances in Neural …, 2023‏ - proceedings.neurips.cc
A unifying theme in the design of intelligent agents is to efficiently optimize a policy based on
what prior knowledge of the problem is available and what actions can be taken to learn …

Collaborative learning of discrete distributions under heterogeneity and communication constraints

X Huang, D Lee, E Dobriban… - Advances in neural …, 2022‏ - proceedings.neurips.cc
In modern machine learning, users often have to collaborate to learn distributions that
generate the data. Communication can be a significant bottleneck. Prior work has studied …

Transfer learning for mortality risk: A case study on the United Kingdom

A Nalmpatian, C Heumann, L Alkaya, W Jackson - medRxiv, 2024‏ - medrxiv.org
This study introduces a transfer learning framework to address data scarcity in mortality risk
prediction for the UK, where local mortality data is unavailable. By leveraging a pretrained …