20,015 Hits in 5.0 sec

Black-Box Optimization with Local Generative Surrogates [article]

Sergey Shirobokov, Vladislav Belavin, Michael Kagan, Andrey Ustyuzhanin, Atılım Güneş Baydin
2020 arXiv   pre-print
We propose a novel method for gradient-based optimization of black-box simulators using differentiable local surrogate models.  ...  We demonstrate that these local surrogates can be used to approximate the gradient of the simulator, and thus enable gradient-based optimization of simulator parameters.  ...  AU and VB are supported by the Russian Science Foundation under grant agreement n • 19-71-30020 for their work on Bayesian Optimization and FFJORD methods.  ... 
arXiv:2002.04632v2 fatcat:ztooib6qqbhe3h2drb6ml2jkbu

Generative Evolutionary Strategy For Black-Box Optimizations [article]

Changhwi Park, Seong Ryeol Kim, Young-Gu Kim, Dae Sin Kim
2022 arXiv   pre-print
This study investigates a novel black-box optimization method based on evolution strategy and generative neural network model.  ...  This hybrid model enables reliable training of surrogate networks; it optimizes multi-objective, high-dimensional, and stochastic black-box functions.  ...  Like Local Generative Surrogates Optimization (L-GSO) [21] , surrogate model-based optimizations are also related to GEO.  ... 
arXiv:2205.03056v1 fatcat:zendzqjkdbavdbikf2qbv4odk4

MATSuMoTo: The MATLAB Surrogate Model Toolbox For Computationally Expensive Black-Box Global Optimization Problems [article]

Juliane Mueller
2014 arXiv   pre-print
MATSuMoTo is the MATLAB Surrogate Model Toolbox for computationally expensive, black-box, global optimization problems that may have continuous, mixed-integer, or pure integer variables.  ...  The multimodality of the expensive black-box function requires an algorithm that is able to search locally as well as globally. MATSuMoTo is able to address these challenges.  ...  Thus, for black-box optimization problems for which it is not known how many local and global optima exist, it is necessary to use a global optimization algorithm that is able to continue the search globally  ... 
arXiv:1404.4261v1 fatcat:gagu23tv7rcizc3oatgsli32i4

Surrogate-based methods for black-box optimization

Ky Khac Vu, Claudia D'Ambrosio, Youssef Hamadi, Leo Liberti
2016 International Transactions in Operational Research  
In this paper, we survey methods that are currently used in black-box optimization, i.e. the kind of problems whose objective functions are very expensive to evaluate and no analytical or derivative information  ...  We concentrate on a particular family of methods, in which surrogate (or meta) models are iteratively constructed and used to search for global solutions.  ...  Heuristical methods There are many difficulties associated with black-box optimization.  ... 
doi:10.1111/itor.12292 fatcat:yoqh5glwrfatzis4kirs55dpeu

Defining Locality for Surrogates in Post-hoc Interpretablity [article]

Thibault Laugel, Xavier Renard, Marie-Jeanne Lesot, Christophe Marsala, Marcin Detyniecki
2018 arXiv   pre-print
Local surrogate models, to approximate the local decision boundary of a black-box classifier, constitute one approach to generate explanations for the rationale behind an individual prediction made by  ...  This paper highlights the importance of defining the right locality, the neighborhood on which a local surrogate is trained, in order to approximate accurately the local black-box decision boundary.  ...  Regarding the surrogate fitting step (2), the surrogate of LIME is trained to approximate locally the black-box decision boundary with a linear regression with regularization (ridge).  ... 
arXiv:1806.07498v1 fatcat:xnuhbkfn5jbkjhca2pol5yyy4m

Global optimization of expensive black-box models based on asynchronous hybrid-criterion with interval reduction [article]

Chunlin Gong, Xu Li, Hua Su, Jinlei Guo, Liangxian Gu
2018 arXiv   pre-print
black-box models.  ...  In this paper, a new sequential surrogate-based optimization (SSBO) algorithm is developed, which aims to improve the global search ability and local search efficiency for the global optimization of expensive  ...  This method is not a general optimization algorithm, while the existing general optimization algorithms are its foundations. SSBO is only applicable to the expensive black-box optimization problem.  ... 
arXiv:1811.12142v1 fatcat:xhsaazqe7zhudioekz3j3pfy2e

Query-Free Adversarial Transfer via Undertrained Surrogates [article]

Chris Miller, Soroush Vosoughi
2020 arXiv   pre-print
We introduce a new method for improving the efficacy of adversarial attacks in a black-box setting by undertraining the surrogate model which the attacks are generated on.  ...  presence of local loss maxima which hinder transferability.  ...  The primary divide is between white box and black-box attacks.  ... 
arXiv:2007.00806v2 fatcat:66okirzm7rabva23eesijzza4u

MISO: mixed-integer surrogate optimization framework

Juliane Müller
2015 Optimization and Engineering  
We introduce MISO, the Mixed-Integer Surrogate Optimization framework. MISO aims at solving computationally expensive black-box optimization problems with mixed-integer variables.  ...  We develop a general surrogate model framework and show how sampling strategies of well-known surrogate model algorithms for continuous optimization can be modified for mixed-integer variables.  ...  Previous Surrogate Model Algorithms for Continuous Optimization Several surrogate model algorithms have been introduced in the literature for addressing computationally expensive black-box optimization  ... 
doi:10.1007/s11081-015-9281-2 fatcat:mfi2xcrjpjhsnhnxlwoqivuxv4

They Might NOT Be Giants: Crafting Black-Box Adversarial Examples with Fewer Queries Using Particle Swarm Optimization [article]

Rayan Mosli, Matthew Wright, Bo Yuan, Yin Pan
2019 arXiv   pre-print
In this paper, we present AdversarialPSO, a black-box attack that uses fewer queries to create adversarial examples with high success rates.  ...  For black-box attacks, the only capability available to the attacker is the ability to query the target with specially crafted inputs and observing the labels returned by the model.  ...  As both attacks require knowledge of model internals-information that is not available in a black-box setting-the authors resort to a local white-box surrogate that approximates the black-box target.  ... 
arXiv:1909.07490v1 fatcat:nye2is5h2zd4tlo44ea42iw7yu

An Improved Surrogate Based Optimization Method for Expensive Black-box Problems

Pengcheng Ye, Guang Pan
2019 IOP Conference Series: Materials Science and Engineering  
For expensive black-box problems, surrogate modelling techniques are generally used to decrease the computational source.  ...  In this study, an improved surrogate based optimization (SBO) method is presented to solve the real-world engineering applications with expensive black-box objective responses.  ...  In ISGO-HSR method, the ensemble of surrogates is used to decrease the required evaluations of the expensive black-box functions.  ... 
doi:10.1088/1757-899x/646/1/012030 fatcat:55nle4d6i5dv3dahxmv62bzhiy

Select Wisely and Explain: Active Learning and Probabilistic Local Post-hoc Explainability [article]

Aditya Saini, Ranjitha Prasad
2022 arXiv   pre-print
We show that UnRAvEL is an efficient surrogate dataset generator by deriving importance scores on this surrogate dataset using sparse linear models.  ...  We present a theoretical analysis of UnRAvEL by treating it as a local optimizer and analyzing its regret in terms of instantaneous regrets over a global optimizer.  ...  As mentioned earlier, in BO, we optimize a black-box objective function that is expensive to evaluate.  ... 
arXiv:2108.06907v2 fatcat:gwkawsvm2fdexa4kc5ui6bgrhi

High-dimensional Black-box Optimization Under Uncertainty [article]

Hadis Anahideh, Jay Rosenberger, Victoria Chen
2021 arXiv   pre-print
Optimizing expensive black-box systems with limited data is an extremely challenging problem.  ...  As a resolution, we present a new surrogate optimization approach by addressing two gaps in prior research – unimportant input variables and inefficient treatment of uncertainty associated with the black-box  ...  General Surrogate Optimization Framework Surrogate optimization is a derivative-free optimization technique, which applies cheap to evaluate surrogate models to optimize expensive black-box functions.  ... 
arXiv:1911.02457v5 fatcat:jqewpgfvsnfmtlte4el4u5ei5q

Framework for Embedding Process Simulator in GAMS via Kriging Surrogate Model Applied to C3MR Natural Gas Liquefaction Optimization

Lucas F. Santos, Caliane B. B. Costa, José A. Caballero, Mauro A. S. S. Ravagnani
2021 Chemical Engineering Transactions  
Rigorous black-box simulations are useful to describe complex systems.  ...  In the present paper, a framework is proposed to embed the Aspen HYSYS process simulator in GAMS using kriging surrogate models to replace the simulator-dependent, black-box objective, and constraints  ...  , and/or black-box functions in black-box optimization problems.  ... 
doi:10.3303/cet2188079 doaj:a2cf2cbb6d084f41b3e148f77c6e6437 fatcat:sklquyp6kjfc5fk22g232scxj4

Empirical Evaluation of Contextual Policy Search with a Comparison-based Surrogate Model and Active Covariance Matrix Adaptation [article]

Alexander Fabisch
2018 arXiv   pre-print
It is based on the standard black-box optimization algorithm CMA-ES.  ...  We will show that improvements with these methods can be impressive in terms of sample-efficiency, although this is not relevant any more for the robotic domain.  ...  Contextual black-box optimization can benefit from the ideas of black-box optimization as contextual policy search can benefit from policy search.  ... 
arXiv:1810.11491v1 fatcat:ipz7yc6i4bedpazsvjdxkkyobq

Surrogate-Based Black-Box Optimization Method for Costly Molecular Properties [article]

Jules Leguy, Thomas Cauchy, Beatrice Duval, Benoit Da Mota
2021 arXiv   pre-print
We propose here a surrogate-based black box optimization method, to tackle jointly the optimization and machine learning problems.  ...  The surrogate is defined as a Gaussian Process Regression (GPR) model, learned on a relevant area of the search space with respect to the property to be optimized.  ...  Preliminaries Surrogate-based black-box optimization Black-box optimization (BBO) is the field of study of the optimization of functions whose derivatives are either non-existent or not available.  ... 
arXiv:2110.03522v1 fatcat:weniw3jhcbhh3cdvfflttdmrhu
« Previous Showing results 1 — 15 out of 20,015 results