Local Superior Soups: A Catalyst for Model Merging in Cross-Silo Federated Learning

arXiv:2410.23660 · cs.LG · Submitted 2024-10-31 · Read on arXiv

cs.LG

Submitted: 2024-10-31

Updated: 2026-10-01

Comments: Accepted at NeurIPS 2024

Code: https://github.com/ubc-tea/Local-Superior-Soups

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Adaptation — 3× in this paper · explained in 1 episode
• Federated Learning — 2× in this paper · explained in 17 episodes
• Federated Learning (FL) — 1× in this paper · explained in 7 episodes
• Model Merging — 1× in this paper · explained in 2 episodes

Sources

• Federated Asymptotics: a model to compare federated learning algorithms
• Where to start? Analyzing the potential value of intermediate models
• One-Shot Federated Learning
• When Do Flat Minima Optimizers Work?
• Scaling Laws for Neural Language Models
• Deep Model Fusion: A Survey
• Experiments with Rich Regime Training for Deep Learning
• Affinity and Diversity: Quantifying Mechanisms of Data Augmentation
• Where to Begin? On the Impact of Pre-Training and Initialization in Federated Learning
• Model Ratatouille: Recycling Diverse Models for Out-of-Distribution Generalization
• Aya Dataset: An Open-Access Collection for Multilingual Instruction Tuning
• Towards Personalized Federated Learning
• LLaMA: Open and Efficient Foundation Language Models
• Llama 2: Open Foundation and Fine-Tuned Chat Models
• A Field Guide to Federated Optimization
• Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms
• FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks