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 2108.01219math.OCcs.LGcs.SCcs.SYeess.SY
收藏

计算牛顿步长比Hessian累加快

Computing the Newton-step faster than Hessian accumulation

SonyAI(索尼人工智能);University of Washington(华盛顿大学)

Akshay Srinivasan, Emanuel Todorov

摘要:

使用$N$决策变量计算通用函数的牛顿步长需要$O(N^3)$触发器。在本文中,我们表明,给定的计算图的功能,这个界可以减少到$O(mτ^3)$,其中$τ, m$是树分解的图的宽度和大小。该算法将基于LQR的非线性最优控制方法推广到一般的最优化问题,即使在Hessian密集的情况下,也能在迭代复杂度上获得非平凡的增益。

英文摘要:

Computing the Newton-step of a generic function with $N$ decision variables takes $O(N^3)$ flops. In this paper, we show that given the computational graph of the function, this bound can be reduced to $O(mτ^3)$, where $τ, m$ are the width and size of a tree-decomposition of the graph. The proposed algorithm generalizes nonlinear optimal-control methods based on LQR to general optimization problems and provides non-trivial gains in iteration-complexity even in cases where the Hessian is dense.

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。

↑