# Machine Unlearning on CIFAR-10 (Utility and Specific Privacy Metrics)

Source: [Machine Unlearning on CIFAR-10 (Utility and Specific Privacy Metrics)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics>)

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

Data updated: 2026-06-01T07:54:21.482Z

Evaluation of Machine Unlearning effectiveness on the CIFAR-10 dataset, focusing on utility retention metrics (Acc\_f, Acc\_r, Acc\_ft, Acc\_rt) alongside specific privacy/unlearning quality indicators like OU @ epsilon and Proto-Acc\_f.

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 | Machine Unlearning |
| Dataset (including any stated split) | CIFAR-10 |
| Metric | Accuracy (Forget) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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 (Forget)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Accuracy%20(Forget)>): Higher is better; unit: PERCENT.

- [Accuracy (Retain)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Accuracy%20(Retain)>): Higher is better; unit: PERCENT.

- [Accuracy (Full Test)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Accuracy%20(Full%20Test)>): Higher is better; unit: PERCENT.

- [Accuracy (Retest)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Accuracy%20(Retest)>): Higher is better; unit: PERCENT.

- [OU @ epsilon](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=OU%20%40%20epsilon>): Higher is better; unit: SCALAR.

- [Prototype Accuracy (Forget)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Prototype%20Accuracy%20(Forget)>): Higher is better; unit: PERCENT.

- [Accuracy\* (Retain)](<https://www.sota2.com/research/sota/machine-unlearning-on-cifar-10-utility-and-specific-privacy-metrics.md?metric=Accuracy*%20(Retain)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Forget) | Result date | Report paper |
| --- | --- | --- | --- |
| Original Model | 87.92 | Not recorded | 68fa79171e209131cea7c144 |
| Original Model | 86.62 | Not recorded | 68fa79171e209131cea7c144 |
| Spotter | 0.5 | Not recorded | 68fa79171e209131cea7c144 |
| Spotter | 0.3 | Not recorded | 68fa79171e209131cea7c144 |
| Retrain Model | 0 | Not recorded | 68fa79171e209131cea7c144 |
| Retrain Model | 0 | Not recorded | 68fa79171e209131cea7c144 |

### Original Model

Accuracy (Forget): 87.92. Result ID: 6a1d38defbc373592c161e9b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"DeiT-S"}. This record does not assert independent reproduction.

### Original Model

Accuracy (Forget): 86.62. Result ID: 6a1d38defbc373592c161e98.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"ViT-S"}. This record does not assert independent reproduction.

### Spotter

Accuracy (Forget): 0.5. Result ID: 6a1d38defbc373592c161e9d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"DeiT-S"}. This record does not assert independent reproduction.

### Spotter

Accuracy (Forget): 0.3. Result ID: 6a1d38defbc373592c161e9a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"ViT-S"}. This record does not assert independent reproduction.

### Retrain Model

Accuracy (Forget): 0. Result ID: 6a1d38defbc373592c161e99.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"ViT-S"}. This record does not assert independent reproduction.

### Retrain Model

Accuracy (Forget): 0. Result ID: 6a1d38defbc373592c161e9c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79171e209131cea7c144.md>)

Recorded evaluation configuration: {"Model":"DeiT-S"}. This record does not assert independent reproduction.

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