Apr 13, 2011 · We explore hybrid methods that exhibit the benefits of both approaches. Rate-of-convergence analysis shows that by controlling the sample size ...

The basic idea of this algorithm is to incorporate block coordinate decomposition and an incremental block averaging scheme into the classic (stochastic) mirror ...

Many structured data-fitting applications require the solution of an optimization problem involving a sum over a potentially large number of measurements.

Many structured data-fitting applications require the solution of an optimization problem involving a sum over a potentially large number of measurements.

It is shown that the inner product test improves upon the well known norm test, and can be used as a basis for an algorithm that is globally convergent on ...

We have generalized our analysis to a variety of scenarios: Newton-like scaling of the gradient (next section). Convex (but not necessarily strongly convex) ...

Apr 12, 2011 · We explore hybrid methods that exhibit the benefits of both approaches. Rate of convergence analysis and numerical experiments illustrate the ...

Many structured data-fitting applications require the solution of an optimization problem involving a sum over a potentially large number of measurements.

Jan 10, 2012 · In this plot we see that the stochastic method, with a carefully chosen step size, can outperform both the deterministic and hybrid methods. We ...

Apr 23, 2020 · Michael P. Friedlander, Mark Schmidt: Hybrid Deterministic-Stochastic Methods for Data Fitting. CoRR abs/1104.2373 (2011).