Explicit Gradient Learning for Black-Box Optimization

Elad Sarafian, Mor Sinay, Yoram Louzoun, Noa Agmon, Sarit Kraus
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:8480-8490, 2020.

Abstract

Black-Box Optimization (BBO) methods can find optimal policies for systems that interact with complex environments with no analytical representation. As such, they are of interest in many Artificial Intelligence (AI) domains. Yet classical BBO methods fall short in high-dimensional non-convex problems. They are thus often overlooked in real-world AI tasks. Here we present a BBO method, termed Explicit Gradient Learning (EGL), that is designed to optimize high-dimensional ill-behaved functions. We derive EGL by finding weak spots in methods that fit the objective function with a parametric Neural Network (NN) model and obtain the gradient signal by calculating the parametric gradient. Instead of fitting the function, EGL trains a NN to estimate the objective gradient directly. We prove the convergence of EGL to a stationary point and its robustness in the optimization of integrable functions. We evaluate EGL and achieve state-of-the-art results in two challenging problems: (1) the COCO test suite against an assortment of standard BBO methods; and (2) in a high-dimensional non-convex image generation task.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-sarafian20a, title = {Explicit Gradient Learning for Black-Box Optimization}, author = {Sarafian, Elad and Sinay, Mor and Louzoun, Yoram and Agmon, Noa and Kraus, Sarit}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {8480--8490}, year = {2020}, editor = {Hal Daumé III and Aarti Singh}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/sarafian20a/sarafian20a.pdf}, url = { http://proceedings.mlr.press/v119/sarafian20a.html }, abstract = {Black-Box Optimization (BBO) methods can find optimal policies for systems that interact with complex environments with no analytical representation. As such, they are of interest in many Artificial Intelligence (AI) domains. Yet classical BBO methods fall short in high-dimensional non-convex problems. They are thus often overlooked in real-world AI tasks. Here we present a BBO method, termed Explicit Gradient Learning (EGL), that is designed to optimize high-dimensional ill-behaved functions. We derive EGL by finding weak spots in methods that fit the objective function with a parametric Neural Network (NN) model and obtain the gradient signal by calculating the parametric gradient. Instead of fitting the function, EGL trains a NN to estimate the objective gradient directly. We prove the convergence of EGL to a stationary point and its robustness in the optimization of integrable functions. We evaluate EGL and achieve state-of-the-art results in two challenging problems: (1) the COCO test suite against an assortment of standard BBO methods; and (2) in a high-dimensional non-convex image generation task.} }
Endnote
%0 Conference Paper %T Explicit Gradient Learning for Black-Box Optimization %A Elad Sarafian %A Mor Sinay %A Yoram Louzoun %A Noa Agmon %A Sarit Kraus %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-sarafian20a %I PMLR %P 8480--8490 %U http://proceedings.mlr.press/v119/sarafian20a.html %V 119 %X Black-Box Optimization (BBO) methods can find optimal policies for systems that interact with complex environments with no analytical representation. As such, they are of interest in many Artificial Intelligence (AI) domains. Yet classical BBO methods fall short in high-dimensional non-convex problems. They are thus often overlooked in real-world AI tasks. Here we present a BBO method, termed Explicit Gradient Learning (EGL), that is designed to optimize high-dimensional ill-behaved functions. We derive EGL by finding weak spots in methods that fit the objective function with a parametric Neural Network (NN) model and obtain the gradient signal by calculating the parametric gradient. Instead of fitting the function, EGL trains a NN to estimate the objective gradient directly. We prove the convergence of EGL to a stationary point and its robustness in the optimization of integrable functions. We evaluate EGL and achieve state-of-the-art results in two challenging problems: (1) the COCO test suite against an assortment of standard BBO methods; and (2) in a high-dimensional non-convex image generation task.
APA
Sarafian, E., Sinay, M., Louzoun, Y., Agmon, N. & Kraus, S.. (2020). Explicit Gradient Learning for Black-Box Optimization. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:8480-8490 Available from http://proceedings.mlr.press/v119/sarafian20a.html .

Related Material