Specifically, we maintain and store a set of device-shared domain knowledge vectors, which accumulates the knowledge learned from all devices during their lifelong adaptation process.

We introduce a natural and mild restriction on the meta-level combination functions of the local trials.

TPT does not explicitly align the pre-trained CLIP to become aware of the test sample distribution. For the effective test-time adaptation of V -L foundation models, it is crucial to bridge the distribution gap between the pre-training dataset and the downstream evaluation set for high zero-shot generalization.

Here, we present novel adaptations for convolutional neural networks to demonstrate that they are indeed able to process functions as inputs and outputs.