Skip to content
Usually take ERM → easily to get trapped into sharp minimizers.
Minimize over loss f.
Hyper-parameters.Visualize loss-surfaceNon-IID levelDifferent step-sizes + linear decay strategy.Domain generalization gaps (Mean/Variance Gaps).Visualize loss surfaceLack of data and style.
Style-based Learning → Style information.Share style information with other clients.AdaIN → MixStyle → Feature level style sampling
FDG
Copy of Copy of Table
Paper Name
Categories
Type
Recent issues
Target
Motivations
Contributions
Evaluation List
Conf./Jour.
Year
Link
Generalized Federated Learning via Sharpness Aware Minimization
Loss Function Design
Theoretical
General task
FDG
ICML
2022
Federated Domain Generalization with Generalization Adjustment.
Loss Function Design
Optimizer
FDG
General task
Theoretical
Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning.
Data Augmentation
Domain-invariant features
Self-supervised Learning
Engineering
FDG
General task
StableFDG: Style and Attention Based Learning for Federated Domain Generalization
Mixing
(Why get style = Gaussian is legit?).
DiPrompT: Disentangled Prompt Tuning for
Multiple Latent Domain Generalization in Federated Learning
There are no rows in this table
Want to print your doc?
This is not the way.
This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
CtrlP
) instead.