ResearchBenchmarksImage Classification on CIFAR-10 IR=100 long-tail (test) (Adversarial Robustness)For AI assistantsMarkdownContent indexImage Classification on CIFAR-10 IR=100 long-tail (test) (Adversarial Robustness)Follow68.07Clean AccuracyRoBal44.77450.82256.8762.918Apr 6, 2021Nov 29, 2021Jul 25, 2022Mar 20, 2023Nov 12, 2023Jul 7, 2024Mar 2, 2025Evaluation resultsMethodMethodLinksClean AccuracyFGSM AccuracyPGD-20 AccuracyCW AccuracyAA AccuracyMIM AccuracyRoBal2021.0468.07—30.3530.5528.9731.25AT2021.0456.72—27.2727.825.9727.87AT-BSLBackbone=ResNet-18, IR...Backbone=ResNet-18, IR=1002025.0354.7322.5119.1717.3418.29—TAETBackbone=ResNet-18, IR...Backbone=ResNet-18, IR=1002025.0352.6828.6826.5342.322.98—TRADES2021.0445.67—26.9726.5325.9327.29