AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

Towards Federated Foundation Models: Scalable Dataset Pipelines for Group-Structured Learning

Zachary Charles, Nicole Elyse Mitchell, Krishna Pillutla, Michael Reneer, Zachary Garrett

NeurIPS 2023 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Vertical federated learning · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Replace in-memory format with streaming format for large datasets (ablation)Dataset Grouper · FedBookCOtime to iterate (s): Out of memory -> 192Improved
Replace hierarchical format with streaming format (ablation)Dataset Grouper · FedCCnewstime to iterate (s): >7200 -> 248Improved
Use in-memory format instead of streaming on small data (ablation)Dataset Grouper · CIFAR-100time to iterate (s): 9.88 -> 0.0783Improved
Increase cohort size from 8 to 32FedAvg with Dataset Grouper · FedC4data iteration share of round time (%): 7.78 -> 9.33Mixed
Apply learning rate schedules (warmup with decay) at the server (ablation)FedSGD · FedC4training lossImproved
Apply learning rate schedules at the server (ablation)FedAvg · FedC4training lossNo clear effect
Use FedSGD instead of FedAvgserver training · FedC4 validationpre-personalization median loss: 5.64 -> 4.93Improved
Use FedAvg instead of FedSGDserver training · FedC4 validationpost-personalization median loss: 3.38 -> 0.012Improved
Use FedAvg instead of FedSGDserver training · FedBookCOpost-personalization loss: 4.0 -> 2.9Improved
Use FedAvg instead of FedSGDserver training · FedBookCOpre-personalization loss: 4.3 -> 5.0Hurt
Scale the model to 1B parameters with 4 batches per client (ablation)FedAvg and FedSGD · FedC4 validationpre-personalization lossImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy