Summary of An Empirical Study Into Clustering Of Unseen Datasets with Self-supervised Encoders, by Scott C. Lowe et al.
An Empirical Study into Clustering of Unseen Datasets with Self-Supervised Encoders
by Scott C. Lowe, Joakim Bruslund Haurum, Sageev Oore, Thomas B. Moeslund, Graham W. Taylor
First submitted to arxiv on: 4 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper investigates whether pretrained image models can generalize to new datasets without retraining. It deploys pretrained models on unseen datasets and examines whether their embeddings form meaningful clusters. The study uses a suite of benchmarking experiments, employing encoders pretrained solely on ImageNet-1k with either supervised or self-supervised training techniques, deployed on image datasets not seen during training, and clustered with conventional clustering algorithms. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Can you generalize trained models to new data without retraining? Researchers explored this question by using pretrained image models on unseen datasets. They looked at whether the model’s internal “features” (or embeddings) group similar images together. The study tested different types of training and saw how well they worked on new data. Surprisingly, some methods did better than others depending on whether they were used within or outside their original training domain. |
Keywords
» Artificial intelligence » Clustering » Self supervised » Supervised
