# Federated Learning on MNIST (non-IID (Dirichlet))

Source: [Federated Learning on MNIST (non-IID (Dirichlet))](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:07:58.331Z

Evaluation of Federated Learning performance on the MNIST dataset configured with a non-IID distribution generated via Dirichlet sampling, focusing on agent utility metrics.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Federated Learning |
| Dataset (including any stated split) | MNIST (non-IID (Dirichlet)) |
| Metric | Utility (Agent 0) |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 2 of 2 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Utility (Agent 0)](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Utility%20(Agent%200)>): Higher is better; unit: SCALAR.

- [Utility (Agent 1)](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Utility%20(Agent%201)>): Higher is better; unit: SCALAR.

- [Utility (Agent 2)](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Utility%20(Agent%202)>): Higher is better; unit: SCALAR.

- [Average Utility](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Average%20Utility>): Higher is better; unit: SCALAR.

- [Multi Utility](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Multi%20Utility>): Higher is better; unit: SCALAR.

- [Relative Utility Ratio](<https://www.sota2.com/research/sota/federated-learning-on-mnist-non-iid-dirichlet.md?metric=Relative%20Utility%20Ratio>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Utility (Agent 0) | Result date | Report paper |
| --- | --- | --- | --- |
| CoreFed | 0.36 | Not recorded | 68fa6dde1e209131cea3c94b |
| FedAvg | 0.34 | Not recorded | 68fa6dde1e209131cea3c94b |

### CoreFed

Utility (Agent 0): 0.36. Result ID: 698ff98a3a6c314849fd2ea5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6dde1e209131cea3c94b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6dde1e209131cea3c94b.md>)

Recorded evaluation configuration: {"Model":"CNN","Data distribution":"Dirichlet","Loss function":"Cross entropy loss","GPU":"1080 Ti"}. This record does not assert independent reproduction.

### FedAvg

Utility (Agent 0): 0.34. Result ID: 698ff98a3a6c314849fd2ea4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6dde1e209131cea3c94b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e761e209131cea40072.md>)

Recorded evaluation configuration: {"Model":"CNN","Data distribution":"Dirichlet","Loss function":"Cross entropy loss","GPU":"1080 Ti"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.