Stochastic compositional gradient descent: algorithms for minimizing compositions of expected-value functions

Mengdi Wang, Ethan X. Fang, Han Liu

Research output: Contribution to journalArticlepeer-review

129 Scopus citations


Classical stochastic gradient methods are well suited for minimizing expected-value objective functions. However, they do not apply to the minimization of a nonlinear function involving expected values or a composition of two expected-value functions, i.e., the problem min xEv[fv(Ew[ gw(x)]) ]. In order to solve this stochastic composition problem, we propose a class of stochastic compositional gradient descent (SCGD) algorithms that can be viewed as stochastic versions of quasi-gradient method. SCGD update the solutions based on noisy sample gradients of fv, gw and use an auxiliary variable to track the unknown quantity Ew[gw(x) ]. We prove that the SCGD converge almost surely to an optimal solution for convex optimization problems, as long as such a solution exists. The convergence involves the interplay of two iterations with different time scales. For nonsmooth convex problems, the SCGD achieves a convergence rate of O(k-1 / 4) in the general case and O(k-2 / 3) in the strongly convex case, after taking k samples. For smooth convex problems, the SCGD can be accelerated to converge at a rate of O(k-2 / 7) in the general case and O(k-4 / 5) in the strongly convex case. For nonconvex problems, we prove that any limit point generated by SCGD is a stationary point, for which we also provide the convergence rate analysis. Indeed, the stochastic setting where one wants to optimize compositions of expected-value functions is very common in practice. The proposed SCGD methods find wide applications in learning, estimation, dynamic programming, etc.

Original languageEnglish (US)
Pages (from-to)419-449
Number of pages31
JournalMathematical Programming
Issue number1-2
StatePublished - Jan 1 2017

All Science Journal Classification (ASJC) codes

  • Software
  • General Mathematics


Dive into the research topics of 'Stochastic compositional gradient descent: algorithms for minimizing compositions of expected-value functions'. Together they form a unique fingerprint.

Cite this