
A Bayesian framework that effectively integrates prior knowledge from various pre-trained vision and language models to achieve superior low-shot image classification performance with well-calibrated uncertainty estimates.


coremsg

leveraging-gaussian-processes-to-assemble-pre-trained-vision-and-language-models-for-robust-low-shot-image-classification


Leveraging Gaussian Processes to Assemble Pre-trained Vision and Language Models for Robust Low-shot Image Classification
