Collaborative Prediction: To Join or To Disjoin Datasets

Kyung Rok Kim, Yansong Wang, Xiaocheng Li, Guanting Chen

Keywords: Machine Learning, Multi-task Learning, Federated Learning

TL;DR: It introduces a data-driven algorithm to combine datasets and get lower population loss with theoretical guarantees

Abstract: With the recent rise of generative Artificial Intelligence (AI), the need of selecting high-quality dataset to improve machine learning models has garnered increasing attention. However, some part of this topic remains underexplored, even for simple prediction models. In this work, we study the problem of developing practical algorithms that select appropriate dataset to minimize population loss of our prediction model with high probability. Broadly speaking, we investigate when datasets from different sources can be effectively merged to enhance the predictive model's performance, and propose a practical algorithm with theoretical guarantees. By leveraging an oracle inequality and data-driven estimators, the algorithm reduces population loss with high probability. Numerical experiments demonstrate its effectiveness in both standard linear regression and broader deep learning applications.

Code Link: https://github.com/kkrokii/collaborative_prediction

Readers: auai.org/UAI/2025/Conference, auai.org/UAI/2025/Conference/Area_Chairs, auai.org/UAI/2025/Conference/Reviewers, auai.org/UAI/2025/Conference/Submission679/Authors, auai.org/UAI/2025/Conference/Submission679/Reproducibility_Reviewers

Submission Number: 679