We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL)
methods: the pre-trained knowledge is indeed a confounder that limits the performance. This …

With the development of deep networks on various large-scale datasets, a large zoo of
pretrained models are available. When transferring from a model zoo, applying classic …

Dataset replication is a useful tool for assessing whether improvements in test accuracy on a
specific benchmark correspond to improvements in models' ability to generalize reliably. In …

Transfer learning aims to leverage knowledge from pre-trained models to benefit the target
task. Prior transfer learning work mainly transfers from a single model. However, with the …

Humans can learn a variety of concepts and skills incrementally over the course of their lives
while exhibiting many desirable properties, such as continual learning without forgetting …

We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL)
methods: the pre-trained knowledge is indeed a confounder that limits the performance. This …

Humans can learn a variety of concepts and skills incrementally over the course of their lives
while exhibiting many desirable properties, such as continual learning without forgetting …