1: \begin{abstract}

2: In this paper, we consider Frank-Wolfe-based algorithms for composite convex optimization problems with objective involving a logarithmically-homogeneous, self-concordant functions. Recent Frank-Wolfe-based methods for this class of problems assume an oracle that returns exact solutions of a linearized subproblem. We relax this assumption and propose a variant of the Frank-Wolfe method with inexact oracle for this class of problems. We show that our inexact variant enjoys similar convergence guarantees to the exact case, while allowing considerably more flexibility in approximately solving the linearized subproblem. In particular, our approach can be applied if the subproblem can be solved prespecified additive error or to prespecified relative error (even though the optimal value of the subproblem may not be uniformly bounded). Furthermore, our approach can also handle the situation where the subproblem is solved via a randomized algorithm that fails with positive probability. Our inexact oracle model is motivated by certain large-scale semidefinite programs where the subproblem reduces to computing an extreme eigenvalue-eigenvector pair, and we demonstrate the practical performance of our algorithm with numerical experiments on problems of this form.   

3: 

4: 

5: % We propose a method when we have access to an oracle that can solve the linear subproblem only up to some prespecified additive or relative error. Furthermore, we also consider a regime where the oracle \textit{fails} to solve the linear subproblem. In this case, we propose a method that, with high probability, generates an approximate solution to the input problem. Our method can be used in large-scale settings where it is more practical to generate and use low-accuracy solutions to the subproblem. We provide provable convergence guarantees for the proposed methods.

6: \end{abstract}

7: