# Image Classification on CelebA Synthetic Patch Neutral Distribution (Dneut) shortcut (blond hair) (test)

Source: [Image Classification on CelebA Synthetic Patch Neutral Distribution (Dneut) shortcut (blond hair) (test)](<https://www.sota2.com/research/sota/image-classification-on-celeba-synthetic-patch-neutral-distribution-dneut-shortcut-blond-hair-test>)

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

Data updated: 2026-02-27T07:43:52.704Z

Evaluation of standard Image Classification performance on the test split of the CelebA Synthetic Patch Neutral Distribution (Dneut) dataset, specifically focusing on the 'shortcut (blond hair)' subset, typically measured by 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 | Image Classification |
| Dataset (including any stated split) | CelebA Synthetic Patch Neutral Distribution (Dneut) shortcut (blond hair) (test) |
| Metric | Accuracy |
| 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](<https://www.sota2.com/research/sota/image-classification-on-celeba-synthetic-patch-neutral-distribution-dneut-shortcut-blond-hair-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Invariant | 94.57 | Not recorded | 68fa6e051e209131cea3d717 |
| JTT | 92.76 | Not recorded | 68fa6e051e209131cea3d717 |
| Chroma-VAE | 92.33 | Not recorded | 68fa6e051e209131cea3d717 |
| Chroma-VAE | 91.82 | Not recorded | 68fa6e051e209131cea3d717 |
| Naive-Independent | 91.49 | Not recorded | 68fa6e051e209131cea3d717 |
| Naive-Class | 90.9 | Not recorded | 68fa6e051e209131cea3d717 |
| Naive-VAE-Class | 89.47 | Not recorded | 68fa6e051e209131cea3d717 |
| Class-Conditional | 61.19 | Not recorded | 68fa6e051e209131cea3d717 |

### Invariant

Accuracy: 94.57. Result ID: 69a14b2f35dfcabb00e61252.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673c1e209131cea15133.md>)

[Implementation](<https://github.com/facebookresearch/InvariantRiskMinimization>)

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

### JTT

Accuracy: 92.76. Result ID: 69a14b2f35dfcabb00e61257.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ade1e209131cea2aeeb.md>)

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

### Chroma-VAE

Accuracy: 92.33. Result ID: 69a14b2f35dfcabb00e61258.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Implementation](<https://github.com/Wanqianxn/chroma-vae-public>)

Recorded evaluation configuration: {"classifier":"z1-clf"}. This record does not assert independent reproduction.

### Chroma-VAE

Accuracy: 91.82. Result ID: 69a14b2f35dfcabb00e61259.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Implementation](<https://github.com/Wanqianxn/chroma-vae-public>)

Recorded evaluation configuration: {"classifier":"z2-clf"}. This record does not assert independent reproduction.

### Naive-Independent

Accuracy: 91.49. Result ID: 69a14b2f35dfcabb00e61255.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

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

### Naive-Class

Accuracy: 90.9. Result ID: 69a14b2f35dfcabb00e61253.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673c1e209131cea15133.md>)

[Implementation](<https://github.com/facebookresearch/InvariantRiskMinimization>)

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

### Naive-VAE-Class

Accuracy: 89.47. Result ID: 69a14b2f35dfcabb00e61254.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e2d1e209131cea3e5fb.md>)

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

### Class-Conditional

Accuracy: 61.19. Result ID: 69a14b2f35dfcabb00e61256.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e051e209131cea3d717.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.