1 paper
Alessio Masano, Giovanni Bellitto, Dipam Goswani +2
Continual learning in transformers is commonly addressed through parameter-efficient adaptation: prompts, adapters, or LoRA modules are specialized per task while the backbone rema…