Paper “Distilling Out-of-Distribution Robustness from Vision-Language Foundation Models” is accepted by NeurIPS 2023.