arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1703.00729cs.LG
收藏

混合复杂性及其在神经网络中的应用

Mixing Complexity and its Applications to Neural Networks

The Hebrew University of Jerusalem(耶路撒冷希伯来大学)

Michal Moshkovitz, Naftali Tishby

摘要:

我们建议通过空间约束的棱镜来分析神经网络。我们观察到,在实践中应用的大多数训练算法使用有界内存,这使我们能够使用一个新的概念,我们称之为混合复杂性的时空权衡的研究。这个概念是为了测量使用有界记忆算法的学习能力而设计的。在本文中,我们描述了我们如何使用混合复杂性,以获得新的结果,什么可以和不能使用神经网络学习。

英文摘要:

We suggest analyzing neural networks through the prism of space constraints. We observe that most training algorithms applied in practice use bounded memory, which enables us to use a new notion introduced in the study of space-time tradeoffs that we call mixing complexity. This notion was devised in order to measure the (in)ability to learn using a bounded-memory algorithm. In this paper we describe how we use mixing complexity to obtain new results on what can and cannot be learned using neural networks.

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。

↑