BriefGPT.xyz
Oct, 2018
利用生成模型进行持续分类学习
Continual Classification Learning Using Generative Models
HTML
PDF
Frantzeska Lavda, Jason Ramapuram, Magda Gregorova, Alexandros Kalousis
TL;DR
本文介绍了一种连续学习的分类模型，可以防止神经网络在逐个学习任务时出现catastrophic forgetting。作者通过扩展已有技术，并提出了一种新的变分下界方法，使得神经网络在连续学习中可以同时保持之前学到的知识。
Abstract
continual learning
is the ability to sequentially learn over time by accommodating knowledge while retaining previously learned experiences.
neural networks
can learn multiple tasks when trained on them jointly,
→