Improving Pre-Trained Self-Supervised Embeddings Through Effective Entropy Maximization

Autor:	Chakraborty, Deep, LeCun, Yann, Rudner, Tim G. J., Learned-Miller, Erik
Rok vydání:	2024
Předmět:	Computer Science - Machine Learning Computer Science - Computer Vision and Pattern Recognition Computer Science - Information Theory Statistics - Applications Statistics - Machine Learning
Druh dokumentu:	Working Paper
Popis:	A number of different architectures and loss functions have been applied to the problem of self-supervised learning (SSL), with the goal of developing embeddings that provide the best possible pre-training for as-yet-unknown, lightly supervised downstream tasks. One of these SSL criteria is to maximize the entropy of a set of embeddings in some compact space. But the goal of maximizing the embedding entropy often depends--whether explicitly or implicitly--upon high dimensional entropy estimates, which typically perform poorly in more than a few dimensions. In this paper, we motivate an effective entropy maximization criterion (E2MC), defined in terms of easy-to-estimate, low-dimensional constraints. We demonstrate that using it to continue training an already-trained SSL model for only a handful of epochs leads to a consistent and, in some cases, significant improvement in downstream performance. We perform careful ablation studies to show that the improved performance is due to the proposed add-on criterion. We also show that continued pre-training with alternative criteria does not lead to notable improvements, and in some cases, even degrades performance. Comment: 19 pages including appendix, 5 figures
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2411.15931 Zobrazit plný text záznamu View this record from Arxiv