First- and Second-Order High Probability Complexity Bounds for Trust-Region Methods with Noisy Oracles

In this paper, we present convergence guarantees for a modified trust-region method designed for minimizing objective functions whose value is computed with noise and for which gradient and Hessian estimates are inexact and possibly random. In order to account for the noise, the method utilizes a relaxed step acceptance criterion and a cautious trust-region radius … Read more

A Theoretical and Empirical Comparison of Gradient Approximations in Derivative-Free Optimization

In this paper, we analyze several methods for approximating gradients of noisy functions using only function values. These methods include finite differences, linear interpolation, Gaussian smoothing and smoothing on a unit sphere. The methods differ in the number of functions sampled, the choice of the sample points, and the way in which the gradient approximations … Read more

Convergence Rate Analysis of a Stochastic Trust Region Method via Supermartingales

We propose a novel framework for analyzing convergence rates of stochastic optimization algorithms with adaptive step sizes. This framework is based on analysing properties of an underlying generic stochastic process, in particular by deriving a bound on the expected stopping time of this process. We utilise this framework to analyse the bounds on expected global … Read more

Optimal Decision Trees for Categorical Data via Integer Programming

Decision trees have been a very popular class of predictive models for decades due to their interpretability and good performance on categorical features. However, they are not always robust and tend to overfit the data. Additionally, if allowed to grow large, they lose interpretability. In this paper, we present a novel mixed integer programming formulation … Read more

A Stochastic Trust Region Algorithm Based on Careful Step Normalization

An algorithm is proposed for solving stochastic and finite sum minimization problems. Based on a trust region methodology, the algorithm employs normalized steps, at least as long as the norms of the stochastic gradient estimates are within a specified interval. The complete algorithm—which dynamically chooses whether or not to employ normalized steps—is proved to have … Read more

Global convergence rate analysis of unconstrained optimization methods based on probabilistic models

We present global convergence rates for a line-search method which is based on random first-order models and directions whose quality is ensured only with certain probability. We show that in terms of the order of the accuracy, the evaluation complexity of such a method is the same as its counterparts that use deterministic accurate models; … Read more

Stochastic Optimization using a Trust-Region Method and Random Models

In this paper, we propose and analyze a trust-region model-based algorithm for solving unconstrained stochastic optimization problems. Our framework utilizes random models of an objective function $f(x)$, obtained from stochastic observations of the function or its gradient. Our method also utilizes estimates of function values to gauge progress that is being made. The convergence analysis … Read more

Alternating direction methods for non convex optimization with applications to second-order least-squares and risk parity portfolio selection

In this paper we mainly focus on optimization of sums of squares of quadratic functions, which we refer to as second-order least-squares problems, subject to convex constraints. Our motivation arises from applications in risk parity portfolio selection. We generalize the setting further by considering a class of nonlinear, non convex functions which admit a (non … Read more

Practical Inexact Proximal Quasi-Newton Method with Global Complexity Analysis

Recently several methods were proposed for sparse optimization which make careful use of second-order information [11, 30, 17, 3] to improve local convergence rates. These methods construct a composite quadratic approximation using Hessian information, optimize this approximation using a first-order method, such as coordinate descent and employ a line search to ensure sufficient descent. Here … Read more

Least-squares approach to risk parity in portfolio selection

The risk parity optimization problem aims to find such portfolios for which the contributions of risk from all assets are equally weighted. Portfolios constructed using risk parity approach are a compromise between two well-known diversification techniques: minimum variance optimization approach and the equal weighting approach. In this paper, we discuss the problem of finding portfolios … Read more