FjORD: Fair and Accurate Federated Learning under heterogeneous targets with Ordered Dropout#

View on GitHub

Paper: openreview.net/forum?id=4fLr7H5D_eT

Authors: Samuel Horváth*, Stefanos Laskaridis*, Mario Almeida*, Ilias Leontiadis, Stylianos Venieris, Nicholas Donald Lane

Abstract: Federated Learning (FL) has been gaining significant traction across different ML tasks, ranging from vision to keyboard predictions. In large-scale deployments, client heterogeneity is a fact and constitutes a primary problem for fairness, training performance and accuracy. Although significant efforts have been made into tackling statistical data heterogeneity, the diversity in the processing capabilities and network bandwidth of clients, termed system heterogeneity, has remained largely unexplored. Current solutions either disregard a large portion of available devices or set a uniform limit on the model’s capacity, restricted by the least capable participants.

In this work, we introduce Ordered Dropout, a mechanism that achieves an ordered, nested representation of knowledge in Neural Networks and enables the extraction of lower footprint submodels without the need for retraining. We further show that for linear maps our Ordered Dropout is equivalent to SVD. We employ this technique, along with a self-distillation methodology, in the realm of FL in a framework called FjORD. FjORD alleviates the problem of client system heterogeneity by tailoring the model width to the client’s capabilities. Extensive evaluation on both CNNs and RNNs across diverse modalities shows that FjORD consistently leads to significant performance gains over state-of-the-art baselines while maintaining its nested structure.

About this baseline#

What’s implemented: The code in this directory implements the two variants of FjORD, with and without knowledge distillation.

Datasets: CIFAR-10

Hardware Setup: We trained the baseline on an Nvidia RTX 4090.

Contributors: @stevelaskaridis (Brave Software), @SamuelHorvath (MBZUAI)

Experimental Setup#

Task: Image Classification

Model: ResNet-18

Dataset:

Feature

Value

Dataset

CIFAR-10

Partition

Randomised Sequential Split

Number of Partitions

100 clients

Data points per client

500 samples

Training Hyperparameters:

Hyperparameter

Value

batch size

32

learning rate

0.1

learning rate scheduler

static

optimiser

sgd

momentum

0

nesterov

False

weight decay

1e-4

sample per round

10

local epochs

1

p-values

[0.2, 0.4, 0.6, 0.8, 1.0]

client tier allocation

uniform

Environment Setup#

Through regular pip#

pip install -r requirements.txt
python setup.py install

Through poetry#

# Set python version
pyenv install 3.10.6
pyenv local 3.10.6

# Tell poetry to use python 3.10
poetry env use 3.10.6

# install the base Poetry environment
poetry install

# activate the environment
poetry shell

Running the Experiments#

Through your environment#

python -m fjord.main  # without knowledge distillation
# or
python -m fjord.main +train_mode=fjord_kd  # with knowledge distillation

Through poetry#

poetry run python -m fjord.main  # without knowledge distillation
# or
poetry run python -m fjord.main +train_mode=fjord_kd  # with knowledge distillation

Expected Results#

cd scripts/
./run.sh

Plots and the associated code reside in fjord/notebooks/visualise.ipynb.

resnet18_cifar10_500_global_rounds_acc_pvalues