# Defense against Data Reconstruction Attack on STL10 (first 100 samples)

Source: [Defense against Data Reconstruction Attack on STL10 (first 100 samples)](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-first-100-samples>)

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

Data updated: 2026-02-26T05:51:03.106Z

Evaluation of defense mechanisms against data reconstruction attacks specifically on the first 100 samples of the STL10 dataset, reporting image quality metrics (MSE, PSNR, SSIM).

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 (first 100 samples) |
| Metric | MSE |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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-first-100-samples.md?metric=MSE>): Lower is better; unit: SCALAR.

- [PSNR](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-first-100-samples.md?metric=PSNR>): Higher is better; unit: SCALAR.

- [SSIM](<https://www.sota2.com/research/sota/defense-against-data-reconstruction-attack-on-stl10-first-100-samples.md?metric=SSIM>): Higher is better; unit: PERCENT.

## Recorded results

| Method | MSE | Result date | Report paper |
| --- | --- | --- | --- |
| DNP | 0.0024 | Not recorded | 69438a9c423ea42265729513 |
| InvL-DNP | 0.012 | Not recorded | 69438a9c423ea42265729513 |
| GNP | 0.06 | Not recorded | 69438a9c423ea42265729513 |
| InvL-GNP | 0.081 | Not recorded | 69438a9c423ea42265729513 |

### DNP

MSE: 0.0024. Result ID: 6999578e5698665ec4c2dc35.

[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: {"δ (noise level)":0.03,"Backbone":"ResNet","Batch size":10,"FC layer size":20}. This record does not assert independent reproduction.

### InvL-DNP

MSE: 0.012. Result ID: 6999578e5698665ec4c2dc36.

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

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

Recorded evaluation configuration: {"δ (noise level)":0.07,"Backbone":"ResNet","Batch size":10,"FC layer size":20}. This record does not assert independent reproduction.

### GNP

MSE: 0.06. Result ID: 6999578e5698665ec4c2dc37.

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

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

Recorded evaluation configuration: {"δ (noise level)":0.001,"Backbone":"ResNet","Batch size":10,"FC layer size":20}. This record does not assert independent reproduction.

### InvL-GNP

MSE: 0.081. Result ID: 6999578e5698665ec4c2dc38.

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

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

Recorded evaluation configuration: {"δ (noise level)":0.005,"Backbone":"ResNet","Batch size":10,"FC layer size":20}. This record does not assert independent reproduction.

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