RBFOpt: an open-source library for black-box optimization with costly function evaluations

Costa, Alberto; Nannicini, Giacomo

doi:10.1007/s12532-018-0144-7

Cited by 139 publications

(176 citation statements)

References 42 publications

Supporting

Mentioning

167

Contrasting

Unclassified

Order By: Relevance

“…Within tasks of the same type, we follow a first come, first served policy. Whenever a task of type 1 is completed, it yields a new interpolation point that is added to the set S. Whenever a task of type 2 is completed, we check if the newly determined search point is to be discarded because of several criteria also employed in the serial version of the optimization algorithm (see [14]), and if the search point is accepted, we queue a task of type 1 to evaluate f at it. An undesirable event can in principle occur if, while f is being evaluated at a point y, the same point y is generated as a search point by concurrent tasks of type 2, and therefore the evaluation of f(y) is performed multiple times.…”

Section: Parallelization Of Rbfoptmentioning

confidence: 99%

An effective algorithm for hyperparameter optimization of neural networks

Diaz¹,

Fokoue-Nkoutche²,

Nannicini³

et al. 2017

IBM J. Res. & Dev.

140

View full text Add to dashboard Cite

A major challenge in designing neural network (NN) systems is to determine the best structure and parameters for the network given the data for the machine learning problem at hand. Examples of parameters are the number of layers and nodes, the learning rates, and the dropout rates. Typically, these parameters are chosen based on heuristic rules and manually fine-tuned, which may be very time-consuming, because evaluating the performance of a single parametrization of the NN may require several hours. This paper addresses the problem of choosing appropriate parameters for the NN by formulating it as a box-constrained mathematical optimization problem, and applying a derivative-free optimization tool that automatically and effectively searches the parameter space. The optimization tool employs a radial basis function model of the objective function (the prediction accuracy of the NN) to accelerate the discovery of configurations yielding high accuracy. Candidate configurations explored by the algorithm are trained to a small number of epochs, and only the most promising candidates receive full training. The performance of the proposed methodology is assessed on benchmark sets and in the context of predicting drug-drug interactions, showing promising results. The optimization tool used in this paper is open-source.

show abstract

Section: Parallelization Of Rbfoptmentioning

confidence: 99%

An effective algorithm for hyperparameter optimization of neural networks

Diaz¹,

Fokoue-Nkoutche²,

Nannicini³

et al. 2017

IBM J. Res. & Dev.

140

View full text Add to dashboard Cite

show abstract

“…An example of two surrogate models that interpolate four data points. The solid‐line model is more likely than the dashed‐line one since it is less bumpy (Costa and Nannicini, ).…”

Section: Surrogate‐based Subproblemsmentioning

confidence: 99%

“…In Jakobsson et al. (), the authors show that the optimal error term e can be obtained by solving an explicit linear system:

(\frac{μ - 1}{μ} I^{n \times n} - B^{T} A B) e = B^{T} A B \tilde{f} .

A similar approach to deal with noise is introduced in Costa and Nannicini (). In this paper, the range within which function values are allowed to vary is required to specify.…”

Section: Surrogate‐based Subproblemsmentioning

confidence: 99%

“…A similar approach to deal with noise is introduced in Costa and Nannicini (2014). In this paper, the range within which function values are allowed to vary is required to specify.…”

Section: Solving Black-box Optimization Problemsmentioning

confidence: 99%

“…(), the problem is solved using an evolutionary algorithm called HYPE. A similar problem in architectural designs (Costa and Nannicini, ) is solved using a radial basis functions (RBFs) based method.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Surrogate‐based methods for black‐box optimization

d'Ambrosio

Hamadi

et al. 2016

Int Trans Operational Res

123

View full text Add to dashboard Cite

show abstract

Solvent extraction process design using deep reinforcement learning

Plathottam

Richey

Curry

et al. 2021

J Adv Manuf & Process

View full text Add to dashboard Cite

Many chemical manufacturing and separations processes like solvent extraction comprise hierarchically complex configurations of functional process units. With increasing complexity, strategies that rely on heuristics become less reliable for design optimization. In this study, we explore deep reinforcement learning for mapping the space of feasible designs to find an optimization strategy that can match or exceed the performance of conventional optimization. To this end, we implement a highly configurable learning environment for the solvent design process to which we can couple state‐of‐the‐art deep reinforcement learning agents. We evaluate the trained agents against the heuristic optimization for the solvent process design tasked to optimize recovery efficiency and product purity. Results demonstrated the agent successfully learned the strategy for predicting comparably optimal solvent extraction process designs for varying combinations of feed compositions.

show abstract

RBFOpt: an open-source library for black-box optimization with costly function evaluations

Cited by 139 publications

References 42 publications

An effective algorithm for hyperparameter optimization of neural networks

An effective algorithm for hyperparameter optimization of neural networks

Surrogate‐based methods for black‐box optimization

Solvent extraction process design using deep reinforcement learning

Contact Info

Product

Resources

About