arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1503.05671cs.LGcs.NEstat.ML
收藏

用Kronecker因子近似曲率优化神经网络

Optimizing Neural Networks with Kronecker-factored Approximate Curvature

University of Toronto(多伦多大学)

James Martens, Roger Grosse

摘要:

我们提出了一种有效的方法来近似自然梯度下降的神经网络,我们称之为克罗内克因子近似曲率(K-FAC)。K-FAC是基于一个有效的可逆近似的神经网络的Fisher信息矩阵,既不是对角也不是低秩,在某些情况下是完全非稀疏的。它是通过近似费舍尔的各种大块(对应于整个层)作为两个小得多的矩阵的克罗内克积而得出的。虽然计算成本仅比普通随机梯度高几倍,但K-FAC产生的更新在优化目标方面取得了更大的进展,这导致实际上可以比具有动量的随机梯度下降快得多的算法。与之前提出的一些使用高质量非对角曲率矩阵(如Hessian-free优化)的近似自然梯度/牛顿方法不同,K-FAC在高度随机的优化机制中工作得非常好。这是因为存储K-FAC近似值并将其求逆到曲率矩阵的成本不取决于用于估计它的数据量,这是通常仅与曲率矩阵的对角或低秩近似相关联的特征。

英文摘要:

We propose an efficient method for approximating natural gradient descent in neural networks which we call Kronecker-Factored Approximate Curvature (K-FAC). K-FAC is based on an efficiently invertible approximation of a neural network's Fisher information matrix which is neither diagonal nor low-rank, and in some cases is completely non-sparse. It is derived by approximating various large blocks of the Fisher (corresponding to entire layers) as being the Kronecker product of two much smaller matrices. While only several times more expensive to compute than the plain stochastic gradient, the updates produced by K-FAC make much more progress optimizing the objective, which results in an algorithm that can be much faster than stochastic gradient descent with momentum in practice. And unlike some previously proposed approximate natural-gradient/Newton methods which use high-quality non-diagonal curvature matrices (such as Hessian-free optimization), K-FAC works very well in highly stochastic optimization regimes. This is because the cost of storing and inverting K-FAC's approximation to the curvature matrix does not depend on the amount of data used to estimate it, which is a feature typically associated only with diagonal or low-rank approximations to the curvature matrix.

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。

↑