Learning generalizable manipulation policy with adapter-based parameter fine-tuning por Lu, K, Ly, KT, Hebberd, W, Zhou, K, Havoutis, I, Markham, A