KITE: A Kernel-based Improved Transferability Estimation Method

arXiv:2405.01603v1 Announce Type: new
Abstract: Transferability estimation has emerged as an important problem in transfer learning. A transferability estimation method takes as inputs a set of pre-trained models and decides which pre-trained model can deliver the best transfer learning performance. Existing methods tackle this problem by analyzing the output of the pre-trained model or by comparing the pre-trained model with a probe model trained on the target dataset. However, neither is sufficient to provide reliable and efficient transferability estimations. In …

abstract arxiv cs.lg inputs kernel performance pre-trained model pre-trained models set transfer transfer learning type