An Empirical Study into Clustering of Unseen Datasets with Self-Supervised Encoders
This work addresses the problem of evaluating model generalization for researchers, but it is incremental as it builds on existing pretraining and clustering methods.
The study investigated whether pretrained image models can generalize to new datasets without retraining by clustering their embeddings, finding that self-supervised encoders prioritize different features than supervised ones and that silhouette scores in UMAP-reduced space correlate with clustering performance.
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 use 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. Supervised encoders typically offer more utility than SSL encoders within the training domain, and vice-versa far outside of it, however, fine-tuned encoders demonstrate the opposite trend. Clustering provides a way to evaluate the utility of self-supervised learned representations orthogonal to existing methods such as kNN. Additionally, we find the silhouette score when measured in a UMAP-reduced space is highly correlated with clustering performance, and can therefore be used as a proxy for clustering performance on data with no ground truth labels. Our code implementation is available at \url{https://github.com/scottclowe/zs-ssl-clustering/}.