# Fair Classification on CelebA (clean)

Source: [Fair Classification on CelebA (clean)](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean>)

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

Data updated: 2026-06-23T06:34:55.344Z

Evaluation of fairness-aware classification performance on the CelebA (clean) dataset, measured using specific fairness metrics like Acc s+ and Accuracy.

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 | Fair Classification |
| Dataset (including any stated split) | CelebA (clean) |
| Metric | Accuracy (s+) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Accuracy (s+)](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean.md?metric=Accuracy%20(s%2B)>): Higher is better; unit: PERCENT.

- [Parity Difference (s+)](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean.md?metric=Parity%20Difference%20(s%2B)>): Lower is better; unit: SCALAR.

- [Accuracy (s-)](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean.md?metric=Accuracy%20(s-)>): Higher is better; unit: PERCENT.

- [Parity Difference (s-)](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean.md?metric=Parity%20Difference%20(s-)>): Lower is better; unit: SCALAR.

- [Accuracy](<https://www.sota2.com/research/sota/fair-classification-on-celeba-clean.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (s+) | Result date | Report paper |
| --- | --- | --- | --- |
| MSAM | 86.32 | Not recorded | 68fa77be1e209131cea750f9 |
| SAM | 85.9 | Not recorded | 68fa77be1e209131cea750f9 |
| FairSAM | 85.74 | Not recorded | 68fa77be1e209131cea750f9 |
| Vanilla | 85.72 | Not recorded | 68fa77be1e209131cea750f9 |
| GroupSAM | 85.71 | Not recorded | 68fa77be1e209131cea750f9 |
| Reweighed | 85.27 | Not recorded | 68fa77be1e209131cea750f9 |
| GroupDRO | 79.79 | Not recorded | 68fa77be1e209131cea750f9 |
| FairReg | 65.3 | Not recorded | 68fa77be1e209131cea750f9 |

### MSAM

Accuracy (s+): 86.32. Result ID: 6a3a2906fbc373592c33ce81.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bc61e209131cea8a5c1.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### SAM

Accuracy (s+): 85.9. Result ID: 6a3a2906fbc373592c33ce80.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69d81e209131cea248fb.md>)

[Implementation](<https://github.com/google-research/sam>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### FairSAM

Accuracy (s+): 85.74. Result ID: 6a3a2906fbc373592c33ce83.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### Vanilla

Accuracy (s+): 85.72. Result ID: 6a3a2906fbc373592c33ce7c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### GroupSAM

Accuracy (s+): 85.71. Result ID: 6a3a2906fbc373592c33ce82.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa705b1e209131cea4a93e.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### Reweighed

Accuracy (s+): 85.27. Result ID: 6a3a2906fbc373592c33ce7e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### GroupDRO

Accuracy (s+): 79.79. Result ID: 6a3a2906fbc373592c33ce7f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67561e209131cea159c7.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

### FairReg

Accuracy (s+): 65.3. Result ID: 6a3a2906fbc373592c33ce7d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77be1e209131cea750f9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63371e209131ce9fd7c7.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-18","Target attribute":"Big Nose","Sensitive attribute":"Age","Test Data":"clean"}. This record does not assert independent reproduction.

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