Optimal Black-Box Reductions Between Optimization Objectives
نویسندگان
چکیده
The diverse world of machine learning applications has given rise to a plethora of algorithms and optimization methods, finely tuned to the specific regression or classification task at hand. We reduce the complexity of algorithm design for machine learning by reductions: we develop reductions that take a method developed for one setting and apply it to the entire spectrum of smoothness and strong-convexity in applications. Furthermore, unlike existing results, our new reductions are optimal and more practical. We show how these new reductions give rise to new and faster running times on training linear classifiers for various families of loss functions, and conclude with experiments showing their successes also in practice.
منابع مشابه
On the Impossibility of Black-Box Transformations in Mechanism Design
We consider the problem of converting an arbitrary approximation algorithm for a singleparameter optimization problem into a computationally efficient truthful mechanism. We ask for reductions that are black-box, meaning that they require only oracle access to the given algorithm and in particular do not require explicit knowledge of the problem constraints. Such a reduction is known to be poss...
متن کاملAn Efficient Pareto Set Identification Approach for Multi-objective Optimization on Black-box Functions
Both multiple objectives and computation-intensive black-box functions often exist simultaneously in engineering design problems. Few of existing multi-objective optimization approaches addresses problems with expensive black-box functions. In this paper, a new method called the Pareto set pursuing (PSP) method is developed. By developing sampling guidance functions based on approximation model...
متن کاملBlack-Box Reductions for Parameter-free Online Learning in Banach Spaces
We introduce several new black-box reductions that significantly improve the design of adaptive and parameterfree online learning algorithms by simplifying analysis, improving regret guarantees, and sometimes even improving runtime. We reduce parameter-free online learning to online exp-concave optimization, we reduce optimization in a Banach space to one-dimensional optimization, and we reduce...
متن کاملOPTIMAL DESIGN OF NON-PRISMATIC REINFORCED CONCRETE BOX GIRDER BRIDGE: MINIMIZATION OF THE COST AND CO2 EMISSION
This paper presents a computational framework for optimal design of non-prismatic reinforced concrete box girder bridges. The variables include the geometry of the cross section, tapered length, concrete strength and reinforcement of box girders and slabs. These are obtained by the enhanced colliding bodies optimization algorithm to optimizing the cost and again CO2 emission. Loading and design...
متن کاملOn Statistical Models Based Multi-objective Optimization
Optimization problems with “expensive” and “black box” objectives are difficult to tackle. Some experience is accumulated in single-objective global optimization of those problems using algorithms based on statistical models of objective functions. The generalization of this approach to the multi-objective optimization is discussed.
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2016