# Defense against Data Reconstruction Attack on STL10 (test)

Source: [Defense against Data Reconstruction Attack on STL10 (test)](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-test>)

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

Data updated: 2026-02-26T05:51:03.091Z

Evaluation of defense mechanisms against data reconstruction attacks on the STL10 test set, reporting image quality metrics (MSE, PSNR, SSIM) and classification accuracy (ACC).

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 | Defense against Data Reconstruction Attack |
| Dataset (including any stated split) | STL10 (test) |
| Metric | MSE |
| Unit | SCALAR |
| Direction | Lower 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

- [MSE](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-test.md?metric=MSE>): Lower is better; unit: SCALAR.

- [PSNR](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-test.md?metric=PSNR>): Higher is better; unit: SCALAR.

- [SSIM](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-test.md?metric=SSIM>): Higher is better; unit: PERCENT.

- [ACC](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-test.md?metric=ACC>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MSE | Result date | Report paper |
| --- | --- | --- | --- |
| Clean | 0.009 | Not recorded | 69438a9c423ea42265729513 |
| DNP | 0.022 | Not recorded | 69438a9c423ea42265729513 |
| Dropout | 0.023 | Not recorded | 69438a9c423ea42265729513 |
| Prune | 0.024 | Not recorded | 69438a9c423ea42265729513 |
| InvL-DNP | 0.024 | Not recorded | 69438a9c423ea42265729513 |
| GNP | 0.039 | Not recorded | 69438a9c423ea42265729513 |
| InvL-GNP | 0.04 | Not recorded | 69438a9c423ea42265729513 |

### Clean

MSE: 0.009. Result ID: 6999578e5698665ec4c2dc2a.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":null}. This record does not assert independent reproduction.

### DNP

MSE: 0.022. Result ID: 6999578e5698665ec4c2dc2d.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c7a1e209131cea342e4.md>)

[Implementation](<https://github.com/zlijingtao/ResSFL>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.07}. This record does not assert independent reproduction.

### Dropout

MSE: 0.023. Result ID: 6999578e5698665ec4c2dc2c.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.98}. This record does not assert independent reproduction.

### Prune

MSE: 0.024. Result ID: 6999578e5698665ec4c2dc2b.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.98}. This record does not assert independent reproduction.

### InvL-DNP

MSE: 0.024. Result ID: 6999578e5698665ec4c2dc2e.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.08}. This record does not assert independent reproduction.

### GNP

MSE: 0.039. Result ID: 6999578e5698665ec4c2dc2f.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e5d1e209131cea3f611.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.5}. This record does not assert independent reproduction.

### InvL-GNP

MSE: 0.04. Result ID: 6999578e5698665ec4c2dc30.

[Reporting paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69438a9c423ea42265729513.md>)

Recorded evaluation configuration: {"Protocol":"Vertical Federated Learning (VFL)","δ / λ":0.6}. This record does not assert independent reproduction.

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