# Image Classification on CIFAR10 (test) (Symmetric/Asymmetric Robustness)

Source: [Image Classification on CIFAR10 (test) (Symmetric/Asymmetric Robustness)](<https://www.sota2.com/research/sota/image-classification-on-cifar10-test-symmetric-asymmetric-robustness>)

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

Data updated: 2026-02-26T06:06:54.138Z

Evaluation of Image Classification performance on the CIFAR-10 test set under specific noise perturbations, measured using Top-1 Accuracy under symmetric and asymmetric conditions (levels 0.2 and 0.4).

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) | CIFAR10 (test) |
| Metric | Top-1 Acc (Sym, 0.2) |
| 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

- [Top-1 Acc (Sym, 0.2)](<https://www.sota2.com/research/sota/image-classification-on-cifar10-test-symmetric-asymmetric-robustness.md?metric=Top-1%20Acc%20(Sym%2C%200.2)>): Higher is better; unit: PERCENT.

- [Top-1 Acc (Sym, 0.4)](<https://www.sota2.com/research/sota/image-classification-on-cifar10-test-symmetric-asymmetric-robustness.md?metric=Top-1%20Acc%20(Sym%2C%200.4)>): Higher is better; unit: PERCENT.

- [Top-1 Acc (Asym, 0.2)](<https://www.sota2.com/research/sota/image-classification-on-cifar10-test-symmetric-asymmetric-robustness.md?metric=Top-1%20Acc%20(Asym%2C%200.2)>): Higher is better; unit: PERCENT.

- [Top-1 Acc (Asym, 0.4)](<https://www.sota2.com/research/sota/image-classification-on-cifar10-test-symmetric-asymmetric-robustness.md?metric=Top-1%20Acc%20(Asym%2C%200.4)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-1 Acc (Sym, 0.2) | Result date | Report paper |
| --- | --- | --- | --- |
| ABSGD(SCE) | 91.15 | Not recorded | 68fa5a7e1e209131ce9c049a |
| CIW(SCE) | 90.79 | Not recorded | 68fa5a7e1e209131ce9c049a |
| ABSGD(CE) | 90.64 | Not recorded | 68fa5a7e1e209131ce9c049a |
| CIW(CE) | 90.29 | Not recorded | 68fa5a7e1e209131ce9c049a |
| EG(SCE) | 90.25 | Not recorded | 68fa5a7e1e209131ce9c049a |
| SGD(SCE) | 90.05 | Not recorded | 68fa5a7e1e209131ce9c049a |
| SGD(TCE) | 89.87 | Not recorded | 68fa5a7e1e209131ce9c049a |
| SGD(CE) | 88.59 | Not recorded | 68fa5a7e1e209131ce9c049a |

### ABSGD(SCE)

Top-1 Acc (Sym, 0.2): 91.15. Result ID: 698c144bf729bc7f68e82af9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Implementation](<https://github.com/qiqi-helloworld/ABSGD/>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"ABSGD","Loss Function":"SCE"}. This record does not assert independent reproduction.

### CIW(SCE)

Top-1 Acc (Sym, 0.2): 90.79. Result ID: 698c144bf729bc7f68e82af7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b371e209131cea2cf9d.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"CIW","Loss Function":"SCE"}. This record does not assert independent reproduction.

### ABSGD(CE)

Top-1 Acc (Sym, 0.2): 90.64. Result ID: 698c144bf729bc7f68e82af8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Implementation](<https://github.com/qiqi-helloworld/ABSGD/>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"ABSGD","Loss Function":"CE"}. This record does not assert independent reproduction.

### CIW(CE)

Top-1 Acc (Sym, 0.2): 90.29. Result ID: 698c144bf729bc7f68e82af6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b371e209131cea2cf9d.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"CIW","Loss Function":"CE"}. This record does not assert independent reproduction.

### EG(SCE)

Top-1 Acc (Sym, 0.2): 90.25. Result ID: 698c144bf729bc7f68e82af5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b41e209131cea239d3.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"EG","Loss Function":"SCE"}. This record does not assert independent reproduction.

### SGD(SCE)

Top-1 Acc (Sym, 0.2): 90.05. Result ID: 698c144bf729bc7f68e82af4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66351e209131cea0ec32.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"SGD","Loss Function":"SCE"}. This record does not assert independent reproduction.

### SGD(TCE)

Top-1 Acc (Sym, 0.2): 89.87. Result ID: 698c144bf729bc7f68e82af3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa651f1e209131cea088eb.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"SGD","Loss Function":"TCE"}. This record does not assert independent reproduction.

### SGD(CE)

Top-1 Acc (Sym, 0.2): 88.59. Result ID: 698c144bf729bc7f68e82af2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa5a7e1e209131ce9c049a.md>)

Recorded evaluation configuration: {"Backbone":"ResNet-50","Pre-trained":"ImageNet","Optimizer":"SGD","Loss Function":"CE"}. This record does not assert independent reproduction.

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