# Machine Unlearning on ImageNet Random Sampling 3-seed average (train)

Source: [Machine Unlearning on ImageNet Random Sampling 3-seed average (train)](<https://www.sota2.com/research/sota/machine-unlearning-on-imagenet-random-sampling-3-seed-average-train>)

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

Data updated: 2026-03-10T09:21:57.839Z

Evaluation of Machine Unlearning performance on ImageNet, specifically using results averaged over 3 random sampling seeds on the training set.

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) | ImageNet Random Sampling 3-seed average (train) |
| Metric | Au Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 13 of 13 |
| 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

- [Au Score](<https://www.sota2.com/research/sota/machine-unlearning-on-imagenet-random-sampling-3-seed-average-train.md?metric=Au%20Score>): Higher is better; unit: SCALAR.

- [Ar Score](<https://www.sota2.com/research/sota/machine-unlearning-on-imagenet-random-sampling-3-seed-average-train.md?metric=Ar%20Score>): Higher is better; unit: SCALAR.

- [Test Accuracy (Atest)](<https://www.sota2.com/research/sota/machine-unlearning-on-imagenet-random-sampling-3-seed-average-train.md?metric=Test%20Accuracy%20(Atest)>): Higher is better; unit: SCALAR.

- [UIS Score](<https://www.sota2.com/research/sota/machine-unlearning-on-imagenet-random-sampling-3-seed-average-train.md?metric=UIS%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Au Score | Result date | Report paper |
| --- | --- | --- | --- |
| SCRUB | 94.54 | Not recorded | 68fa773c1e209131cea71e18 |
| FISHER | 94.36 | Not recorded | 68fa773c1e209131cea71e18 |
| Original | 94.26 | Not recorded | 68fa773c1e209131cea71e18 |
| SSD | 93.5 | Not recorded | 68fa773c1e209131cea71e18 |
| FT | 89.96 | Not recorded | 68fa773c1e209131cea71e18 |
| Influence | 89.08 | Not recorded | 68fa773c1e209131cea71e18 |
| OrthoGrad | 81.04 | Not recorded | 68fa773c1e209131cea71e18 |
| GDR-GMA | 80.41 | Not recorded | 68fa773c1e209131cea71e18 |
| NegGrad+ | 78.75 | Not recorded | 68fa773c1e209131cea71e18 |
| DUCK | 76.75 | Not recorded | 68fa773c1e209131cea71e18 |
| SCAR | 71.49 | Not recorded | 68fa773c1e209131cea71e18 |
| NegGrad | 24.2 | Not recorded | 68fa773c1e209131cea71e18 |
| Retrain | 3.36 | Not recorded | 68fa773c1e209131cea71e18 |

### SCRUB

Au Score: 94.54. Result ID: 69afe2ae6dbb7075b95888de.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71011e209131cea4e7b9.md>)

[Implementation](<https://github.com/Meghdad92/SCRUB>)

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

### FISHER

Au Score: 94.36. Result ID: 69afe2ae6dbb7075b95888dc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67551e209131cea158d5.md>)

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

### Original

Au Score: 94.26. Result ID: 69afe2ae6dbb7075b95888d7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

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

### SSD

Au Score: 93.5. Result ID: 69afe2ae6dbb7075b95888e1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717d1e209131cea51995.md>)

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

### FT

Au Score: 89.96. Result ID: 69afe2ae6dbb7075b95888d9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

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

### Influence

Au Score: 89.08. Result ID: 69afe2ae6dbb7075b95888dd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68f81e209131cea1f643.md>)

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

### OrthoGrad

Au Score: 81.04. Result ID: 69afe2ae6dbb7075b95888e3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

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

### GDR-GMA

Au Score: 80.41. Result ID: 69afe2ae6dbb7075b95888e0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

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

### NegGrad+

Au Score: 78.75. Result ID: 69afe2ae6dbb7075b95888db.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71011e209131cea4e7b9.md>)

[Implementation](<https://github.com/Meghdad92/SCRUB>)

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

### DUCK

Au Score: 76.75. Result ID: 69afe2ae6dbb7075b95888df.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73121e209131cea5a465.md>)

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

### SCAR

Au Score: 71.49. Result ID: 69afe2ae6dbb7075b95888e2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72ee1e209131cea595a3.md>)

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

### NegGrad

Au Score: 24.2. Result ID: 69afe2ae6dbb7075b95888da.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68881e209131cea1cbe9.md>)

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

### Retrain

Au Score: 3.36. Result ID: 69afe2ae6dbb7075b95888d8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa773c1e209131cea71e18.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72cd1e209131cea5888f.md>)

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

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