Preprint · 2019 · Open access

Bayesian Optimization of Composite Functions

Raul Astudillo, Peter I. Frazier

arXiv (Cornell University), 354–363 · Published 4 June 2019 · DOI 10.48550/arxiv.1906.01537

Open PDFView at the publisherCite this paper

In brief

“Bayesian Optimization of Composite Functions” is a 2019 preprint by Raul Astudillo and Peter I. Frazier, published in arXiv (Cornell University). OpenAlex counts 120 citations as of 19 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).

Citations
120
Most cited in
2024 (28)
References
40
Authors
2

Abstract

We consider optimization of composite objective functions, i.e., of the form $f(x)=g(h(x))$, where $h$ is a black-box derivative-free expensive-to-evaluate function with vector-valued outputs, and $g$ is a cheap-to-evaluate real-valued function. While these problems can be solved with standard Bayesian optimization, we propose a novel approach that exploits the composite structure of the objective function to substantially improve sampling efficiency. Our approach models $h$ using a multi-output Gaussian process and chooses where to sample using the expected improvement evaluated on the implied non-Gaussian posterior on $f$, which we call expected improvement for composite functions (\ei). Although \ei\ cannot be computed in closed form, we provide a novel stochastic gradient estimator that allows its efficient maximization. We also show that our approach is asymptotically consistent, i.e., that it recovers a globally optimal solution as sampling effort grows to infinity, generalizing previous convergence results for classical expected improvement. Numerical experiments show that our approach dramatically outperforms standard Bayesian optimization benchmarks, reducing simple regret by several orders of magnitude.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Astudillo, R., & Frazier, P. I. (2019). Bayesian Optimization of Composite Functions. arXiv (Cornell University), 354–363. https://doi.org/10.48550/arxiv.1906.01537

MLA 9

Astudillo, Raul, and Peter I. Frazier. “Bayesian Optimization of Composite Functions.” arXiv (Cornell University), 2019, pp. 354–363. https://doi.org/10.48550/arxiv.1906.01537.

Chicago

Astudillo, Raul and Peter I. Frazier. 2019. “Bayesian Optimization of Composite Functions.” arXiv (Cornell University): 354–363. https://doi.org/10.48550/arxiv.1906.01537.

Harvard

Astudillo, R. and Frazier, P. I. (2019) ‘Bayesian Optimization of Composite Functions’, arXiv (Cornell University), pp. 354–363. doi:10.48550/arxiv.1906.01537.

IEEE

R. Astudillo and P. I. Frazier, “Bayesian Optimization of Composite Functions,” arXiv (Cornell University), pp. 354–363, 2019, doi: 10.48550/arxiv.1906.01537.

BibTeX

@misc{astudillo2019bayesian,
  title     = {{Bayesian Optimization of Composite Functions}},
  author    = {Astudillo, Raul and Frazier, Peter I.},
  howpublished = {arXiv (Cornell University)},
  year      = {2019},
  pages     = {354--363},
  doi       = {10.48550/arxiv.1906.01537},
  url       = {https://doi.org/10.48550/arxiv.1906.01537}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Citations by year

YearCitationsShare of the best year
201955
202066
20211111
202299
20231818
20242828
20251818
20262525

Topic and keywords

Social Sciences › Decision Sciences › Management Science and Operations Research › Advanced Bandit Algorithms Research

  • Bayesian optimization
  • composite objective function
  • multi-output Gaussian processes
  • expected improvement
  • sampling efficiency
  • asymptotic consistency
  • black-box optimization
  • expensive-to-evaluate functions

Work with this paper

Most-cited papers in Decision Sciences

All highly cited Decision Sciences papers

Questions about this paper

How do I cite this paper in APA style?

Astudillo, R., & Frazier, P. I. (2019). Bayesian Optimization of Composite Functions. arXiv (Cornell University), 354–363. https://doi.org/10.48550/arxiv.1906.01537

How many times has this paper been cited?

120 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.

Is this paper free to read?

Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Raul Astudillo, Peter I. Frazier.

What is the DOI of this paper?

10.48550/arxiv.1906.01537, permanently at https://doi.org/10.48550/arxiv.1906.01537. Sign in and open it from this page.

Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.