01 · The problem

The hospital that knows most about rare diseases gets one of the smaller votes.

Standard federated averaging (FedAvg) weights each hospital by how many images it holds in total, not by what it knows. On the real Fed-ISIC2019 split that shortchanges the one hospital that specialises in the rare diseases.
The gap

Centre 2, in one number.

more of the rare-disease data than its share of the vote.

Rare-disease training images held by Centre 227.0%
Aggregation weight FedAvg gives Centre 214.5%

104 of 385 rare-disease training images · 2,691 of 18,597 training images overall

Every hospital

The same comparison, for all six.

Only Centre 2 holds far more of the rare-disease data than its weight. Centre 4 holds none at all - yet under FedAvg it still votes on rare diseases.

Centre 0

12,413 images in total

Rare-disease training images
184
Share of all rare-disease data
47.8%
Share of FedAvg weight
53.4%
Knowledge vs weight
0.90×

Centre 1

3,954 images in total

Rare-disease training images
68
Share of all rare-disease data
17.7%
Share of FedAvg weight
17.0%
Knowledge vs weight
1.04×

Centre 2

Specialist

3,363 images in total

Rare-disease training images
104
Share of all rare-disease data
27.0%
Share of FedAvg weight
14.5%
Knowledge vs weight
1.87×

Centre 3

2,259 images in total

Rare-disease training images
23
Share of all rare-disease data
6.0%
Share of FedAvg weight
9.7%
Knowledge vs weight
0.61×

Centre 4

No rare images

819 images in total

Rare-disease training images
0
Share of all rare-disease data
0.0%
Share of FedAvg weight
3.5%
Knowledge vs weight
0.00×

Centre 5

439 images in total

Rare-disease training images
6
Share of all rare-disease data
1.6%
Share of FedAvg weight
1.9%
Knowledge vs weight
0.83×

Side by side

Training split. Centre 2 highlighted.

What counts as rare

Two diseases, about one percent each.

A disease is rare here if it holds under 1/25 of the largest class's share (Melanocytic nevus, 48.7%) - a 1.95% cutoff. The rule was chosen after seeing the counts, and we say so.
Rare

Dermatofibroma

1.03%

of all images

4 of 6

hospitals hold ≥ 20 training images

Rare

Vascular lesion

1.09%

of all images

3 of 6

hospitals hold ≥ 20 training images

Who can check

Few hospitals can vouch for a rare disease.

Counting hospitals with at least 20 training images of each disease. If one of only three holders misreports, there may be nobody left to contradict it - and it isn't only the rare diseases that are thinly held.
DiseaseShare of dataHospitals holding it
Melanoma18.00%6 of 6
Melanocytic nevus48.72%6 of 6
Basal cell carcinoma14.29%3 of 6
Actinic keratosis3.73%3 of 6
Benign keratosis10.44%5 of 6
DermatofibromaRare1.03%4 of 6
Vascular lesionRare1.09%3 of 6
Squamous cell carcinoma2.70%2 of 6