# Image Classification on Fashion-MNIST (Adversarial & Robustness Metrics)

Source: [Image Classification on Fashion-MNIST (Adversarial & Robustness Metrics)](<https://www.sota2.com/research/sota/image-classification-on-fashion-mnist-adversarial-and-robustness-metrics>)

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

Data updated: 2026-05-11T07:43:33.086Z

Evaluation of Image Classification performance on the Fashion-MNIST dataset, reporting clean accuracy, adversarial accuracy (at epsilon 0.30), AURAC score, and category count.

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 | Image Classification |
| Dataset (including any stated split) | Fashion-MNIST |
| Metric | Clean Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [Clean Accuracy](<https://www.sota2.com/research/sota/image-classification-on-fashion-mnist-adversarial-and-robustness-metrics.md?metric=Clean%20Accuracy>): Higher is better; unit: PERCENT.

- [Adversarial Accuracy (eps=0.30)](<https://www.sota2.com/research/sota/image-classification-on-fashion-mnist-adversarial-and-robustness-metrics.md?metric=Adversarial%20Accuracy%20(eps%3D0.30)>): Higher is better; unit: PERCENT.

- [AURAC Score](<https://www.sota2.com/research/sota/image-classification-on-fashion-mnist-adversarial-and-robustness-metrics.md?metric=AURAC%20Score>): Higher is better; unit: SCALAR.

- [Category Count](<https://www.sota2.com/research/sota/image-classification-on-fashion-mnist-adversarial-and-robustness-metrics.md?metric=Category%20Count>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Clean Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Sep-Aware | 82.9 | Not recorded | 6a01628edbc836b87d8230b7 |
| AdvTrain | 82.8 | Not recorded | 6a01628edbc836b87d8230b7 |
| Selective | 82.4 | Not recorded | 6a01628edbc836b87d8230b7 |
| Two-Stage Sel. | 82.4 | Not recorded | 6a01628edbc836b87d8230b7 |
| AdvTrain | 82.2 | Not recorded | 6a01628edbc836b87d8230b7 |
| Selective | 81.1 | Not recorded | 6a01628edbc836b87d8230b7 |
| Vanilla | 80.7 | Not recorded | 6a01628edbc836b87d8230b7 |

### Sep-Aware

Clean Accuracy: 82.9. Result ID: 6a01888f47f34ce924e30fcc.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### AdvTrain

Clean Accuracy: 82.8. Result ID: 6a01888f47f34ce924e30fc9.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665a1e209131cea0fbae.md>)

Recorded evaluation configuration: {"Adversarial Training":"on"}. This record does not assert independent reproduction.

### Selective

Clean Accuracy: 82.4. Result ID: 6a01888f47f34ce924e30fca.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

Recorded evaluation configuration: {"Selective Mechanism":"off"}. This record does not assert independent reproduction.

### Two-Stage Sel.

Clean Accuracy: 82.4. Result ID: 6a01888f47f34ce924e30fcd.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### AdvTrain

Clean Accuracy: 82.2. Result ID: 6a01888f47f34ce924e30fc8.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665a1e209131cea0fbae.md>)

Recorded evaluation configuration: {"Adversarial Training":"off"}. This record does not assert independent reproduction.

### Selective

Clean Accuracy: 81.1. Result ID: 6a01888f47f34ce924e30fcb.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

Recorded evaluation configuration: {"Selective Mechanism":"on"}. This record does not assert independent reproduction.

### Vanilla

Clean Accuracy: 80.7. Result ID: 6a01888f47f34ce924e30fc7.

[Reporting paper](<https://www.sota2.com/research/paper/6a01628edbc836b87d8230b7.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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