Conditioning by adaptive sampling for robust design

David Brookes, Hahnbeom Park, Jennifer Listgarten
Proceedings of the 36th International Conference on Machine Learning, PMLR 97:773-782, 2019.

Abstract

We present a method for design problems wherein the goal is to maximize or specify the value of one or more properties of interest (e.g. maximizing the fluorescence of a protein). We assume access to black box, stochastic “oracle" predictive functions, each of which maps from design space to a distribution over properties of interest. Because many state-of-the-art predictive models are known to suffer from pathologies, especially for data far from the training distribution, the problem becomes different from directly optimizing the oracles. Herein, we propose a method to solve this problem that uses model-based adaptive sampling to estimate a distribution over the design space, conditioned on the desired properties.

Cite this Paper

BibTeX
@InProceedings{pmlr-v97-brookes19a, title = {Conditioning by adaptive sampling for robust design}, author = {Brookes, David and Park, Hahnbeom and Listgarten, Jennifer}, booktitle = {Proceedings of the 36th International Conference on Machine Learning}, pages = {773--782}, year = {2019}, editor = {Chaudhuri, Kamalika and Salakhutdinov, Ruslan}, volume = {97}, series = {Proceedings of Machine Learning Research}, month = {09--15 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v97/brookes19a/brookes19a.pdf}, url = {https://proceedings.mlr.press/v97/brookes19a.html}, abstract = {We present a method for design problems wherein the goal is to maximize or specify the value of one or more properties of interest (e.g. maximizing the fluorescence of a protein). We assume access to black box, stochastic “oracle" predictive functions, each of which maps from design space to a distribution over properties of interest. Because many state-of-the-art predictive models are known to suffer from pathologies, especially for data far from the training distribution, the problem becomes different from directly optimizing the oracles. Herein, we propose a method to solve this problem that uses model-based adaptive sampling to estimate a distribution over the design space, conditioned on the desired properties.} }
Endnote
%0 Conference Paper %T Conditioning by adaptive sampling for robust design %A David Brookes %A Hahnbeom Park %A Jennifer Listgarten %B Proceedings of the 36th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Kamalika Chaudhuri %E Ruslan Salakhutdinov %F pmlr-v97-brookes19a %I PMLR %P 773--782 %U https://proceedings.mlr.press/v97/brookes19a.html %V 97 %X We present a method for design problems wherein the goal is to maximize or specify the value of one or more properties of interest (e.g. maximizing the fluorescence of a protein). We assume access to black box, stochastic “oracle" predictive functions, each of which maps from design space to a distribution over properties of interest. Because many state-of-the-art predictive models are known to suffer from pathologies, especially for data far from the training distribution, the problem becomes different from directly optimizing the oracles. Herein, we propose a method to solve this problem that uses model-based adaptive sampling to estimate a distribution over the design space, conditioned on the desired properties.
APA
Brookes, D., Park, H. & Listgarten, J.. (2019). Conditioning by adaptive sampling for robust design. Proceedings of the 36th International Conference on Machine Learning, in Proceedings of Machine Learning Research 97:773-782 Available from https://proceedings.mlr.press/v97/brookes19a.html.

Related Material