# Image Classification on CIFAR-100 (IID) (Poisoning Robustness) Source: [Image Classification on CIFAR-100 (IID) (Poisoning Robustness)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:39:17.485Z Evaluation of Image Classification performance on the IID split of the CIFAR-100 dataset, specifically measuring robustness against various types of poisoning attacks (Badnet, DBA, Neurotoxin) and clean accuracy. 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) | CIFAR-100 (IID) | | Metric | Clean Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [Clean Accuracy](): Higher is better; unit: PERCENT. - [Badnet Backdoor Accuracy (r=0.3)](): Higher is better; unit: PERCENT. - [Badnet Backdoor Accuracy (r=0.5)](): Higher is better; unit: PERCENT. - [Badnet Error Rate (r=0.3)](): Lower is better; unit: PERCENT. - [Badnet Error Rate (r=0.5)](): Lower is better; unit: PERCENT. - [DBA Backdoor Accuracy (r=0.3)](): Higher is better; unit: PERCENT. - [DBA Backdoor Accuracy (r=0.5)](): Higher is better; unit: PERCENT. - [DBA Error Rate (r=0.3)](): Lower is better; unit: PERCENT. - [DBA Error Rate (r=0.5)](): Lower is better; unit: PERCENT. - [Neurotoxin Backdoor Accuracy (r=0.3)](): Higher is better; unit: PERCENT. - [Neurotoxin Backdoor Accuracy (r=0.5)](): Higher is better; unit: PERCENT. - [Neurotoxin Error Rate (r=0.3)](): Lower is better; unit: PERCENT. - [Neurotoxin Error Rate (r=0.5)](): Lower is better; unit: PERCENT. - [Average Backdoor Accuracy](): Higher is better; unit: PERCENT. - [Average Error Rate](): Lower is better; unit: PERCENT. ## Recorded results | Method | Clean Accuracy | Result date | Report paper | | --- | --- | --- | --- | | FedAvg | 64.29 | Not recorded | 68fa77831e209131cea73bd9 | | FedAvg\* | 64.29 | Not recorded | 68fa77831e209131cea73bd9 | | Foolsgold | 64.13 | Not recorded | 68fa77831e209131cea73bd9 | | AlignIns | 63.45 | Not recorded | 68fa77831e209131cea73bd9 | | MM | 63.26 | Not recorded | 68fa77831e209131cea73bd9 | | Lockdown | 62.88 | Not recorded | 68fa77831e209131cea73bd9 | | RFA | 53.92 | Not recorded | 68fa77831e209131cea73bd9 | | MKrum | 51.28 | Not recorded | 68fa77831e209131cea73bd9 | | RLR | 44.34 | Not recorded | 68fa77831e209131cea73bd9 | ### FedAvg Clean Accuracy: 64.29. Result ID: 69952c103badfaffe6124838. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### FedAvg\* Clean Accuracy: 64.29. Result ID: 69952c103badfaffe6124839. [Reporting paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### Foolsgold Clean Accuracy: 64.13. Result ID: 69952c103badfaffe612483d. [Reporting paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### AlignIns Clean Accuracy: 63.45. Result ID: 69952c103badfaffe6124840. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### MM Clean Accuracy: 63.26. Result ID: 69952c103badfaffe612483e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### Lockdown Clean Accuracy: 62.88. Result ID: 69952c103badfaffe612483f. [Reporting paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### RFA Clean Accuracy: 53.92. Result ID: 69952c103badfaffe612483b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### MKrum Clean Accuracy: 51.28. Result ID: 69952c103badfaffe612483c. [Reporting paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. ### RLR Clean Accuracy: 44.34. Result ID: 69952c103badfaffe612483a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"VGG9"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.