arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1612.04035cs.LG
收藏

DizzyRNN:为保范反向传播重新参数化循环神经网络

DizzyRNN: Reparameterizing Recurrent Neural Networks for Norm-Preserving Backpropagation

Cornell University(康奈尔大学)

Victor Dorobantu, Per Andre Stromhaug, Jess Renteria

摘要:

消失和爆炸梯度问题是研究得很好的障碍,使得循环神经网络很难学习长期的时间依赖性。我们提出了一种标准循环神经网络的重新参数化,通过Givens旋转以可证明的范数保持方式更新线性变换。此外,我们使用绝对价值函数作为逐元素非线性,以保持整个网络上反向传播信号的范数。我们表明,这种重新参数化减少了参数的数量,并保持相同的算法复杂度作为一个标准的递归神经网络,同时优于标准的递归神经网络与正交初始化和长短期记忆网络的复制问题。

英文摘要:

The vanishing and exploding gradient problems are well-studied obstacles that make it difficult for recurrent neural networks to learn long-term time dependencies. We propose a reparameterization of standard recurrent neural networks to update linear transformations in a provably norm-preserving way through Givens rotations. Additionally, we use the absolute value function as an element-wise non-linearity to preserve the norm of backpropagated signals over the entire network. We show that this reparameterization reduces the number of parameters and maintains the same algorithmic complexity as a standard recurrent neural network, while outperforming standard recurrent neural networks with orthogonal initializations and Long Short-Term Memory networks on the copy problem.

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。

↑