Specifically, we observe that CLIP's visual encoder tends to

During the CL of VLMs, we need not only to prevent the catastrophic forgetting on incrementally learned knowledge but also to preserve the zero-shot ability of VLMs.

Implementation of our proposed method and the datasets are now available1.

Neural Information Processing Systems http://nips.cc/

Unlike the conventional machine learning paradigms where learning is performed on a static dataset, domain incremental learning, i.e., continual learning with evolving domains, hopes to accommodate the model to the dynamically changing data distributions, while retaining the knowledge learned from previous domains [

Grid cells in the medial entorhinal cortex create remarkable periodic maps of explored space during navigation. Recent studies show that they form similar maps of abstract cognitive spaces. Examples of such abstract environments include auditory tone sequences in which the pitch is continuously varied or images in which abstract features are continuously deformed (e.g., a cartoon bird whose legs stretch and shrink).