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 2302.07396cs.LGcond-mat.stat-mechcs.AIq-bio.NC
收藏

卷积酉或正交循环神经网络

Convolutional unitary or orthogonal recurrent neural networks

Rockefeller University(洛克菲勒大学)

Marcelo O. Magnasco

摘要:

循环神经网络非常强大,但很难训练。他们的问题之一是消失梯度问题,即训练信号的传播可能会呈指数衰减,冻结训练。已经提出使用正交或酉矩阵来缓解这个问题,其功率既不爆炸也不衰减,但是它们的计算费用阻碍了它们的使用。在这里,我们证明了在卷积RNN的特定情况下,我们可以定义一个卷积指数,并且这个操作将反对称或反厄米卷积核转换为正交或酉卷积核。我们明确推导出基于FFT的算法来计算内核及其衍生物。因此,对正交变换的该子空间进行参数化的计算复杂性与网络的迭代相同。

英文摘要:

Recurrent neural networks are extremely powerful yet hard to train. One of their issues is the vanishing gradient problem, whereby propagation of training signals may be exponentially attenuated, freezing training. Use of orthogonal or unitary matrices, whose powers neither explode nor decay, has been proposed to mitigate this issue, but their computational expense has hindered their use. Here we show that in the specific case of convolutional RNNs, we can define a convolutional exponential and that this operation transforms antisymmetric or anti-Hermitian convolution kernels into orthogonal or unitary convolution kernels. We explicitly derive FFT-based algorithms to compute the kernels and their derivatives. The computational complexity of parametrizing this subspace of orthogonal transformations is thus the same as the networks' iteration.

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。

↑