We consider an important problem in scientific discovery, identifying sparse governing equations for nonlinear dynamical systems. This involves solving sparse ridge regression problems to provable optimality in order to determine which terms drive the underlying dynamics. We propose a fast algorithm, OKRidge, for sparse ridge regression, using a novel lower bound calculation involving, first, a saddle point formulation, and from there, either solving (i) a linear system or (ii) using an ADMM-based approach, where the proximal operators can be efficiently evaluated by solving another linear system and an isotonic regression problem. We also propose a method to warm-start our solver, which leverages a beam search. Experimentally, our methods attain provable optimality with run times that are orders of magnitude faster than those of the existing MIP formulations solved by the commercial solver Gurobi.
翻译:我们考虑科学发现中的一个重要问题,即确定非线性动态系统的稀疏方程式。这涉及到通过演示哪些项驱动潜在动力学来解决稀疏岭回归问题,以可证明的最优性。我们提出了一种快速算法OKRidge,用于稀疏岭回归,采用新颖的下界计算方法,首先使用鞍点形式,然后通过解决(i)线性系统或(ii)使用基于ADMM方法的方法之一,其中可以通过解决另一个线性系统和保序回归问题有效地计算近端算子。我们还提出了一种方法来热启动我们的解算器,这利用了一个集束搜索。实验上,我们的方法以比商业解算器Gurobi解决的现有MIP公式快几个数量级的运行时间实现了可证明的最优性。