ResearchBenchmarksImage Classification on CIFAR-10 Merge Imbalance (pool)Image Classification on CIFAR-10 Merge Imbalance (pool)Follow86.19Balanced Accuracy (10% cost)PriorAL70.933274.894178.85582.8159May 30, 2026Evaluation resultsMethodMethodLinksBalanced Accuracy (10% cost)Balanced Accuracy (20% cost)PriorALBackbone=CLIP-L14, Fou...Backbone=CLIP-L14, Foundation Model=CLIP-L142026.0586.1986.29RandomBackbone=CLIP-L14, Fou...Backbone=CLIP-L14, Foundation Model=CLIP-L142026.0571.5277.61