CULT：基于典型性的环境检测的连续无监督学习
CULT: Continual Unsupervised Learning with Typicality-Based Environment Detection
摘要：
我们介绍了CULT（连续无监督表示学习与基于典型性的环境检测），一种新的算法，连续无监督学习与变分自动编码器。CULT在VAE的潜在空间中使用一个简单的典型性度量来检测环境中的分布变化，该度量与生成重放和辅助环境分类器结合使用，以限制无监督表示学习中的灾难性遗忘。在我们的实验中，CULT显著优于基线连续无监督学习方法。本文的代码可以在这里找到：https://github.com/oliveradk/cult
英文摘要：
We introduce CULT (Continual Unsupervised Representation Learning with Typicality-Based Environment Detection), a new algorithm for continual unsupervised learning with variational auto-encoders. CULT uses a simple typicality metric in the latent space of a VAE to detect distributional shifts in the environment, which is used in conjunction with generative replay and an auxiliary environmental classifier to limit catastrophic forgetting in unsupervised representation learning. In our experiments, CULT significantly outperforms baseline continual unsupervised learning approaches. Code for this paper can be found here: https://github.com/oliveradk/cult
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
