# Machine Unlearning Utility and Membership Inference on CIFAR-10 (test) Source: [Machine Unlearning Utility and Membership Inference on CIFAR-10 (test)]() Collection index: [llms.txt]() Data updated: 2026-06-04T09:57:32.666Z Evaluation of Machine Unlearning effectiveness on the test split of CIFAR-10, reporting aggregate utility metrics (UA, RA, TA) alongside Membership Inference Attack (MIA) performance. 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 (test) | | Metric | Unlearning Accuracy (UA) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [Unlearning Accuracy (UA)](): Higher is better; unit: PERCENT. - [Retention Accuracy (RA)](): Higher is better; unit: PERCENT. - [Target Accuracy (TA)](): Higher is better; unit: PERCENT. - [MIA Success Rate](): Higher is better; unit: PERCENT. - [Average Gap](): Lower is better; unit: PERCENT. - [RTE](): Lower is better; unit: SCALAR. - [MIA False Alarm Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Unlearning Accuracy (UA) | Result date | Report paper | | --- | --- | --- | --- | | Retrain | 100 | Not recorded | 6a210690c10484bac23f0184 | | l1-sparse | 100 | Not recorded | 6a210690c10484bac23f0184 | | GA | 99.9 | Not recorded | 6a210690c10484bac23f0184 | | SalUn | 99.3 | Not recorded | 6a210690c10484bac23f0184 | | IU | 97 | Not recorded | 6a210690c10484bac23f0184 | | PURGE | 91.6 | Not recorded | 6a210690c10484bac23f0184 | | RL | 89.3 | Not recorded | 6a210690c10484bac23f0184 | | FT | 31.7 | Not recorded | 6a210690c10484bac23f0184 | | SFRon | 14.7 | Not recorded | 68fa789e1e209131cea795fa | | RT | 11.17 | Not recorded | 68fa789e1e209131cea795fa | | MCUβ | 8.04 | Not recorded | 68fa789e1e209131cea795fa | | MCU | 7.21 | Not recorded | 68fa789e1e209131cea795fa | | NegGrad+ | 5.22 | Not recorded | 68fa789e1e209131cea795fa | | SalUn | 3.87 | Not recorded | 68fa789e1e209131cea795fa | | NegTV | 3.33 | Not recorded | 68fa789e1e209131cea795fa | | RL | 3.24 | Not recorded | 68fa789e1e209131cea795fa | | FT | 0.34 | Not recorded | 68fa789e1e209131cea795fa | | GA | 0.03 | Not recorded | 68fa789e1e209131cea795fa | ### Retrain Unlearning Accuracy (UA): 100. Result ID: 6a214bfefbc373592c1d6be0. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### l1-sparse Unlearning Accuracy (UA): 100. Result ID: 6a214bfefbc373592c1d6be2. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### GA Unlearning Accuracy (UA): 99.9. Result ID: 6a214bfefbc373592c1d6be6. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### SalUn Unlearning Accuracy (UA): 99.3. Result ID: 6a214bfefbc373592c1d6be1. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### IU Unlearning Accuracy (UA): 97. Result ID: 6a214bfefbc373592c1d6be3. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### PURGE Unlearning Accuracy (UA): 91.6. Result ID: 6a214bfefbc373592c1d6be7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"forget\_class":"0","objective":"kl\_retain"}. This record does not assert independent reproduction. ### RL Unlearning Accuracy (UA): 89.3. Result ID: 6a214bfefbc373592c1d6be4. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### FT Unlearning Accuracy (UA): 31.7. Result ID: 6a214bfefbc373592c1d6be5. [Reporting paper]() Recorded evaluation configuration: {"forget\_class":"0"}. This record does not assert independent reproduction. ### SFRon Unlearning Accuracy (UA): 14.7. Result ID: 6a0450f647f34ce924eaca32. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### RT Unlearning Accuracy (UA): 11.17. Result ID: 6a0450f647f34ce924eaca2d. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### MCUβ Unlearning Accuracy (UA): 8.04. Result ID: 6a0450f647f34ce924eaca36. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### MCU Unlearning Accuracy (UA): 7.21. Result ID: 6a0450f647f34ce924eaca35. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### NegGrad+ Unlearning Accuracy (UA): 5.22. Result ID: 6a0450f647f34ce924eaca31. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### SalUn Unlearning Accuracy (UA): 3.87. Result ID: 6a0450f647f34ce924eaca33. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### NegTV Unlearning Accuracy (UA): 3.33. Result ID: 6a0450f647f34ce924eaca34. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### RL Unlearning Accuracy (UA): 3.24. Result ID: 6a0450f647f34ce924eaca2f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### FT Unlearning Accuracy (UA): 0.34. Result ID: 6a0450f647f34ce924eaca2e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. ### GA Unlearning Accuracy (UA): 0.03. Result ID: 6a0450f647f34ce924eaca30. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"PreResNet-110","Forgetting Ratio":"20%"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.