Frank–Wolfe and friends: a journey into projection-free first-order optimization methods
نویسندگان
چکیده
Abstract Invented some 65 years ago in a seminal paper by Marguerite Straus-Frank and Philip Wolfe, the Frank–Wolfe method recently enjoys remarkable revival, fuelled need of fast reliable first-order optimization methods Data Science other relevant application areas. This review tries to explain success this approach illustrating versatility applicability wide range contexts, combined with an account on recent progress variants, improving both speed efficiency surprisingly simple principle optimization.
منابع مشابه
First-order Methods for Geodesically Convex Optimization
Geodesic convexity generalizes the notion of (vector space) convexity to nonlinear metric spaces. But unlike convex optimization, geodesically convex (g-convex) optimization is much less developed. In this paper we contribute to the understanding of g-convex optimization by developing iteration complexity analysis for several first-order algorithms on Hadamard manifolds. Specifically, we prove ...
متن کاملStochastic First-Order Methods with Random Constraint Projection
We consider convex optimization problems with structures that are suitable for sequential treatment or online sampling. In particular, we focus on problems where the objective function is an expected value, and the constraint set is the intersection of a large number of simpler sets. We propose an algorithmic framework for stochastic first-order methods using random projection/proximal updates ...
متن کاملMultilevel Projection Methods for First - Order SystemLeast Squares
This paper focuses on the multilevel projection method (PML) applied to numerical solution of the basic equations of uid dynamics formulated as least-squares problems for rst-order systems. The fundamental approach taken here is to pose the uid ow equations in their rst-order form, incorporate additional (usually redundant) equations where they improve the character of the formulation, and deen...
متن کاملtragic contradictions: a comparative study of characterization in eugene o’neill’s long day’s journey into night and mahmud dowlatabadi’s tangna
در طی چند دهه ی اخیر، مفهوم «تراژدی» و «قهرمان تراژدی» توجهی روزافرون را تقریباً در تمام حوزه های نقد ادبی به خود معطوف کرده است. برخی نظیر ارسطو، نیچه، و آرتور میلر به بازخوانی آن پرداخته و برخی دیگر نظیر سارتر، استریندبرگ، یوجین اُنیل، برتولت برشت، و آنتونین آرتود به افزودن ابعاد نوینی به این مبحث همت گماشته اند. آنچه قهرمان تراژدی مدرن را از مفهوم کلاسیک آن متمایز می کند نه لغزش تراژیک متداول ...
Proximal and First-Order Methods for Convex Optimization
We describe the proximal method for minimization of convex functions. We review classical results, recent extensions, and interpretations of the proximal method that work in online and stochastic optimization settings.
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: 4OR
سال: 2021
ISSN: ['1614-2411', '1619-4500']
DOI: https://doi.org/10.1007/s10288-021-00493-y