New Uniform Herding method improves class-incremental learning

Researchers have introduced Uniform Herding, a novel method for exemplar replay in class-incremental learning. This technique aims to preserve performance on earlier classes as feature representations evolve by managing a bounded active exemplar set and refreshing chosen exemplars from a candidate pool. Experiments on CIFAR-100 using a ResNet-18 backbone demonstrated that Uniform Herding achieved higher final average accuracy and reduced forgetting compared to the iCaRL method. AI

IMPACT This research could lead to more effective AI systems that can learn new information without forgetting previously acquired knowledge.

RANK_REASON The cluster contains an academic paper detailing a new method for class-incremental learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• CIFAR-100
• ResNet-18
• Uniform Herding

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →