Timezone: »
"Bayesian Optimization is a very effective tool for optimizing expensive black-box functions. Inspired by applications developing and characterizing reaction chemistry using droplet microfluidic reactors, we consider a novel setting where the expense of evaluating the function can increase significantly when making large input changes between iterations. We further assume we are working asynchronously, meaning we have to decide on new queries before we finish evaluating previous experiments. This paper investigates the problem and introduces 'Sequential Bayesian Optimization via Adaptive Connecting Samples' (SnAKe), which provides a solution by considering large batches of queries and preemptively building optimization paths that minimize input costs. We investigate some convergence properties and empirically show that the algorithm is able to achieve regret similar to classical Bayesian Optimization algorithms in both the synchronous and asynchronous settings, while reducing the input costs significantly. We show the method is robust to the choice of its single hyper-parameter and provide a parameter-free alternative."
Author Information
Jose Pablo Folch (Imperial College London)
PhD Student at Imperial College London, as part of the StatML CDT. Working on chemistry-inspired Bayesian Optimization problems.
Shiqiang Zhang (Imperial College London)
Robert Lee (BASF SE)
Behrang Shafei (BASF)
David Walz (BASF SE)
Calvin Tsay (Imperial College London)
Mark van der Wilk (Imperial College London)
Ruth Misener (Imperial College London)
Dr Ruth Misener is Professor in Computational Optimization in the Imperial College London Department of Computing. Ruth received an SB from MIT and a PhD from Princeton. Foundations of her research are in numerical optimization algorithms. Applications include decision-making under uncertainty, energy efficiency, process network design & operations, and scheduling. Ruth's research team makes their software contributions available open source (<https://github.com/cog-imperial>). Ruth received the 2017 Macfarlane Medal from the Royal Academy of Engineering and the 2020 Outstanding Young Researcher Award from the AIChE Computing & Systems Technology Division. She has won best paper awards at AAMAS (2020, Best Innovative Demo), CPAIOR (2021), and the Journal of Global Optimization (2013). Ruth is an associate editor for *Computers & Chemical Engineering* and *INFORMS Journal on Computing*.
More from the Same Authors
-
2022 Poster: Joint Entropy Search for Multi-Objective Bayesian Optimization »
Ben Tu · Axel Gandy · Nikolas Kantas · Behrang Shafei -
2022 : Actually Sparse Variational Gaussian Processes »
Jake Cunningham · So Takao · Mark van der Wilk · Marc Deisenroth -
2022 : Recommendations for Baselines and Benchmarking Approximate Gaussian Processes »
Sebastian Ober · David Burt · Artem Artemev · Mark van der Wilk -
2022 : Sparse Convolutions on Lie Groups »
Tycho van der Ouderaa · Mark van der Wilk -
2022 : Causal Discovery using Marginal Likelihood »
Anish Dhir · Mark van der Wilk -
2022 Poster: Invariance Learning in Deep Neural Networks with Differentiable Laplace Approximations »
Alexander Immer · Tycho van der Ouderaa · Gunnar Rätsch · Vincent Fortuin · Mark van der Wilk -
2022 Poster: Memory safe computations with XLA compiler »
Artem Artemev · Yuze An · Tilman Roeder · Mark van der Wilk -
2022 Poster: Relaxing Equivariance Constraints with Non-stationary Continuous Filters »
Tycho van der Ouderaa · David W. Romero · Mark van der Wilk -
2022 Poster: Tree ensemble kernels for Bayesian optimization with known constraints over mixed-feature spaces »
Alexander Thebelt · Calvin Tsay · Robert Lee · Nathan Sudermann-Merx · David Walz · Behrang Shafei · Ruth Misener -
2021 Poster: Partition-Based Formulations for Mixed-Integer Optimization of Trained ReLU Neural Networks »
Calvin Tsay · Jan Kronqvist · Alexander Thebelt · Ruth Misener -
2020 Poster: A Bayesian Perspective on Training Speed and Model Selection »
Clare Lyle · Lisa Schut · Robin Ru · Yarin Gal · Mark van der Wilk -
2020 Poster: Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertainty »
Miguel Monteiro · Loic Le Folgoc · Daniel Coelho de Castro · Nick Pawlowski · Bernardo Marques · Konstantinos Kamnitsas · Mark van der Wilk · Ben Glocker