Deciding How to Decide: Dynamic Routing in Artificial Neural Networks

Deciding How to Decide: Dynamic Routing in Artificial Neural Networks
复制标题

决定如何决定:人工神经网络中的动态路由

DOI:
--
复制
发表时间:
2017
期刊:
International Conference on Machine Learning
影响因子:
--
通讯作者:
P. Perona
P. Perona
中科院分区:
--
文献类型:
--
作者:
Mason McGill;P. Perona

文献摘要

参考文献

被引文献

相似文献

我们提出并系统地评估了三种用于训练动态路由人工神经网络的策略:学习变换图,通过这些图,不同的输入信号可能采取不同的路径。尽管一些方法比其他方法更有优势,但由此产生的网络在性质上往往是相似的。我们发现,在被训练为对图像进行分类的动态路由网络中,层和分支变得专门处理不同类别的图像。此外,在固定计算预算的情况下,动态路由网络往往比同等的静态路由网络性能更好。
We propose and systematically evaluate three strategies for training dynamically-routed artificial neural networks: graphs of learned transformations through which different input signals may take different paths. Though some approaches have advantages over others, the resulting networks are often qualitatively similar. We find that, in dynamically-routed networks trained to classify images, layers and branches become specialized to process distinct categories of images. Additionally, given a fixed computational budget, dynamically-routed networks tend to perform better than comparable statically-routed networks.
深度(呃)学习。
DOI: 10.1523/jneurosci.0153-18.2018
发表时间: 2018
期刊: The Journal of neuroscience : the official journal of the Society for Neuroscience
影响因子: --
作者:
Srinivasan,Shyam;Greenspan,RalphJ;Stevens,CharlesF;Grover,Dhruv
通讯作者: Grover,Dhruv