Paper 解读 - Arxiv 论文中文解读平台 - GptGet

排序: 最新热门引用

全部 Artificial Intelligence Computation and Language Computer Vision Information Retrieval Machine Learning Machine Learning (Stats) Neural and Evolutionary Computing Robotics

cs.NE 1603.08983

Adaptive Computation Time for Recurrent Neural Networks

提出自适应计算时间（ACT）算法，动态调节RNN的计算步数，显著提升多任务性能。

Alex Graves

2016-03-30 34

cs.NE 1506.02626

Learning both Weights and Connections for Efficient Neural Networks

提出一种剪枝方法，通过学习重要连接，AlexNet参数缩减9倍，VGG-16缩减13倍，保持准确率。

Song Han, Jeff Pool, John Tran 等

2015-06-09 35

cs.NE 1404.5997

One weird trick for parallelizing convolutional neural networks

提出一种结合数据并行与模型并行的卷积神经网络训练加速方法，显著提升GPU扩展性。

Alex Krizhevsky

2014-04-24 57