מחקר אמפירי בנושא קבוצתיות של נתונים לא ראויים עם קודדים עצמאיים

תקציר מקורי באנגליתarXiv:2406.02465v2 Announce Type: replace Abstract: Can pretrained models generalize to new datasets without any retraining? We deploy pretrained image models on datasets they were not trained for, and investigate whether their embeddings form meaningful clusters. Our suite of benchmarking experiments uses encoders pretrained solely on ImageNet-1k with either supervised or self-supervised training techniques, deployed on image datasets that were not seen during training, and clustered with conventional clustering algorithms. This evaluation provides new insights into the embeddings of self-supervised models, which prioritize different features to supervised models. We find evidence that supervised encoders offer more utility than SSL encoders within the training domain, and vice-versa far