Privately Customizing Prefinetuning to Better Match User Data in Federated Learning

Charlie Hou, Hongyuan Zhan, Akshat Shrivastava, Sid Wang, Aleksandr Livshits, Giulia Fanti, Daniel Lazar

Keywords: federated learning, differential privacy, pretraining, prefinetuning, large language models

TL;DR: We demonstrate a proof-of-concept for a new approach in private FL training: (1) customize a prefinetuning dataset to better match user data (2) prefinetune (3) perform FL-finetuning.

Abstract: In Federated Learning (FL), accessing private client data incurs communication and privacy costs. As a result, FL deployments commonly prefinetune pretrained foundation models on a (large, possibly public) dataset that is held by the central server; they then FL-finetune the model on a private, federated dataset held by clients. Evaluating prefinetuning dataset quality reliably and privately (with respect to its usefulness on the user datasets) is therefore of high importance. To this end, we propose FreD (Federated Private Fréchet Distance) --- a privately computed distance between a prefinetuning dataset and federated datasets. Intuitively, it privately computes and compares a Fréchet distance between embeddings generated by a large language model on both the central (public) dataset and the federated private client data. To make this computation privacy-preserving, we use distributed, differentially-private mean and covariance estimators. We show empirically that FreD accurately predicts the best prefinetuning dataset at minimal privacy cost. Altogether, using FreD we demonstrate a proof-of-concept for a new approach in private FL training: (1) customize a prefinetuning dataset to better match user data (2) prefinetune (3) perform FL-finetuning.