arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2004.06479math.OC
收藏

快速随机拟牛顿法

Faster Stochastic Quasi-Newton Methods

Qingsong Zhang, Feihu Huang, Cheng Deng, Heng Huang

摘要:

随机最优化方法已经成为机器学习中一类流行的优化工具。特别地,随机梯度下降(SGD)由于低的每次迭代计算复杂性而被广泛用于机器学习问题,例如训练神经网络。事实上,利用二阶信息的牛顿或拟牛顿法能够获得比一阶方法更好的解。因此,随机拟牛顿(SQN)方法已经发展到实现更好的解决方案,有效地利用近似二阶信息比随机一阶方法。然而,现有的SQN方法仍然没有达到最著名的随机一阶预言(SFO)的复杂性。为了填补这一空白,我们提出了一种新的更快的随机拟牛顿法(SpiderSQN)的方差减少技术的SIPDER的基础上。我们证明了我们的SpiderSQN方法达到了最好的已知的SFO复杂性的$\mathcal{O}(n+n^{1/2}ε^{-2})$在有限和设置,以获得一个$ε$ -一阶稳定点。为了进一步提高其实际性能,我们将SpiderSQN与不同的动量方案相结合。此外,将提出的算法推广到在线环境,得到了$\mathcal{O}(ε^{-3})$算法的SFO复杂度,与已有的最优结果一致.在基准数据集上进行的大量实验表明,我们的新算法在非凸优化方面优于最先进的方法。

英文摘要:

Stochastic optimization methods have become a class of popular optimization tools in machine learning. Especially, stochastic gradient descent (SGD) has been widely used for machine learning problems such as training neural networks due to low per-iteration computational complexity. In fact, the Newton or quasi-newton methods leveraging second-order information are able to achieve a better solution than the first-order methods. Thus, stochastic quasi-Newton (SQN) methods have been developed to achieve the better solution efficiently than the stochastic first-order methods by utilizing approximate second-order information. However, the existing SQN methods still do not reach the best known stochastic first-order oracle (SFO) complexity. To fill this gap, we propose a novel faster stochastic quasi-Newton method (SpiderSQN) based on the variance reduced technique of SIPDER. We prove that our SpiderSQN method reaches the best known SFO complexity of $\mathcal{O}(n+n^{1/2}ε^{-2})$ in the finite-sum setting to obtain an $ε$-first-order stationary point. To further improve its practical performance, we incorporate SpiderSQN with different momentum schemes. Moreover, the proposed algorithms are generalized to the online setting, and the corresponding SFO complexity of $\mathcal{O}(ε^{-3})$ is developed, which also matches the existing best result. Extensive experiments on benchmark datasets demonstrate that our new algorithms outperform state-of-the-art approaches for nonconvex optimization.

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。

↑