Decoupled and distilled: Task-adaptive LoRA-teachers with ensemble knowledge transfer for few-shot class-incremental learning
Few-Shot Class-Incremental Learning (FSCIL) addresses the challenge of learning new classes from very limited samples while retaining knowledge of previously learned ones. Although parameter-efficient fine-tuning methods with pre-trained models show promise for class-incremental learning, strict gradient-based constrai...
Hong-Wei Zhao, Rui Liu, Yan-Song Liu et al.
· Neurocomputing · 0 citations