-
Boundary Exploration for Bayesian Optimization With Unknown Physical Constraints
Authors:
Yunsheng Tian,
Ane Zuniga,
Xinwei Zhang,
Johannes P. Dürholt,
Payel Das,
Jie Chen,
Wojciech Matusik,
Mina Konaković Luković
Abstract:
Bayesian optimization has been successfully applied to optimize black-box functions where the number of evaluations is severely limited. However, in many real-world applications, it is hard or impossible to know in advance which designs are feasible due to some physical or system limitations. These issues lead to an even more challenging problem of optimizing an unknown function with unknown const…
▽ More
Bayesian optimization has been successfully applied to optimize black-box functions where the number of evaluations is severely limited. However, in many real-world applications, it is hard or impossible to know in advance which designs are feasible due to some physical or system limitations. These issues lead to an even more challenging problem of optimizing an unknown function with unknown constraints. In this paper, we observe that in such scenarios optimal solution typically lies on the boundary between feasible and infeasible regions of the design space, making it considerably more difficult than that with interior optima. Inspired by this observation, we propose BE-CBO, a new Bayesian optimization method that efficiently explores the boundary between feasible and infeasible designs. To identify the boundary, we learn the constraints with an ensemble of neural networks that outperform the standard Gaussian Processes for capturing complex boundaries. Our method demonstrates superior performance against state-of-the-art methods through comprehensive experiments on synthetic and real-world benchmarks. Code available at: https://github.com/yunshengtian/BE-CBO
△ Less
Submitted 21 May, 2024; v1 submitted 12 February, 2024;
originally announced February 2024.
-
Learning a reactive potential for silica-water through uncertainty attribution
Authors:
Swagata Roy,
Johannes P. Dürholt,
Thomas S. Asche,
Federico Zipoli,
Rafael Gómez-Bombarelli
Abstract:
The reactivity of silicates in an aqueous solution is relevant to various chemistries ranging from silicate minerals in geology, to the C-S-H phase in cement, nanoporous zeolite catalysts, or highly porous precipitated silica. While simulations of chemical reactions can provide insight at the molecular level, balancing accuracy and scale in reactive simulations in the condensed phase is a challeng…
▽ More
The reactivity of silicates in an aqueous solution is relevant to various chemistries ranging from silicate minerals in geology, to the C-S-H phase in cement, nanoporous zeolite catalysts, or highly porous precipitated silica. While simulations of chemical reactions can provide insight at the molecular level, balancing accuracy and scale in reactive simulations in the condensed phase is a challenge. Here, we demonstrate how a machine-learning reactive interatomic potential can accurately capture silicate-water reactivity. The model was trained on a new dataset comprising 400,000 energies and forces of molecular clusters at the $ω$-B97XD def2-TVZP level. To ensure the robustness of the model, we introduce a new and general active learning strategy based on the attribution of the model uncertainty, that automatically isolates uncertain regions of bulk simulations to be calculated as small-sized clusters. Our trained potential is found to reproduce static and dynamic properties of liquid water and solid crystalline silicates, despite having been trained exclusively on cluster data. Furthermore, we utilize enhanced sampling simulations to recover the self-ionization reactivity of water accurately, and the acidity of silicate oligomers, and lastly study the silicate dimerization reaction in a water solution at neutral conditions and find that the reaction occurs through a flanking mechanism.
△ Less
Submitted 4 July, 2023;
originally announced July 2023.
-
GAUCHE: A Library for Gaussian Processes in Chemistry
Authors:
Ryan-Rhys Griffiths,
Leo Klarner,
Henry B. Moss,
Aditya Ravuri,
Sang Truong,
Samuel Stanton,
Gary Tom,
Bojana Rankovic,
Yuanqi Du,
Arian Jamasb,
Aryan Deshwal,
Julius Schwartz,
Austin Tripp,
Gregory Kell,
Simon Frieder,
Anthony Bourached,
Alex Chan,
Jacob Moss,
Chengzhi Guo,
Johannes Durholt,
Saudamini Chaurasia,
Felix Strieth-Kalthoff,
Alpha A. Lee,
Bingqing Cheng,
Alán Aspuru-Guzik
, et al. (2 additional authors not shown)
Abstract:
We introduce GAUCHE, a library for GAUssian processes in CHEmistry. Gaussian processes have long been a cornerstone of probabilistic machine learning, affording particular advantages for uncertainty quantification and Bayesian optimisation. Extending Gaussian processes to chemical representations, however, is nontrivial, necessitating kernels defined over structured inputs such as graphs, strings…
▽ More
We introduce GAUCHE, a library for GAUssian processes in CHEmistry. Gaussian processes have long been a cornerstone of probabilistic machine learning, affording particular advantages for uncertainty quantification and Bayesian optimisation. Extending Gaussian processes to chemical representations, however, is nontrivial, necessitating kernels defined over structured inputs such as graphs, strings and bit vectors. By defining such kernels in GAUCHE, we seek to open the door to powerful tools for uncertainty quantification and Bayesian optimisation in chemistry. Motivated by scenarios frequently encountered in experimental chemistry, we showcase applications for GAUCHE in molecular discovery and chemical reaction optimisation. The codebase is made available at https://github.com/leojklarner/gauche
△ Less
Submitted 21 February, 2023; v1 submitted 6 December, 2022;
originally announced December 2022.
-
Identifying the Bottleneck for Heat Transport in Metal-Organic Frameworks
Authors:
Sandro Wieser,
Tomas Kamencek,
Johannes P. Dürholt,
Rochus Schmid,
NataliaBedoya-Martínez,
Egbert Zojer
Abstract:
Controlling the transport of thermal energy is key to most applications of metal-organic frameworks. Analyzing the evolution of the effective local temperature, the interfaces between the metal nodes and the organic linkers are identified as the primary bottlenecks for heat conduction. Consequently, changing the bonding strength at that node-linker interface and the mass of the metal atoms can be…
▽ More
Controlling the transport of thermal energy is key to most applications of metal-organic frameworks. Analyzing the evolution of the effective local temperature, the interfaces between the metal nodes and the organic linkers are identified as the primary bottlenecks for heat conduction. Consequently, changing the bonding strength at that node-linker interface and the mass of the metal atoms can be exploited to tune the thermal conductivity. This insight is generated employing molecular dynamics simulations in conjunction with advanced, ab initio parametrized force fields. The focus of the present study is on MOF-5 as a prototypical example of an isoreticular MOF. Still, the key findings prevail for different node structures and node-linker bonding chemistries. The presented results lay the foundation for develo** detailed structure-to-property relationships for thermal transport in MOFs with the goal of devising strategies for the application-specific optimization of heat conduction.
△ Less
Submitted 24 August, 2020;
originally announced August 2020.
-
Evaluating Computational Shortcuts in Supercell-Based Phonon Calculations of Molecular Crystals: The Instructive Case of Naphthalene
Authors:
Tomas Kamencek,
Sandro Wieser,
Hirotaka Kojima,
Natalia Bedoya-Martínez,
Johannes P. Dürholt,
Rochus Schmid,
Egbert Zojer
Abstract:
Phonons crucially impact a variety of properties of organic semiconductor materials. For instance, charge- and heat transport depend on low-frequency phonons, while for other properties, such as the free energy, especially high-frequency phonons count. For all these quantities one needs to know the entire phonon band structure, whose simulation becomes exceedingly expensive for more complex system…
▽ More
Phonons crucially impact a variety of properties of organic semiconductor materials. For instance, charge- and heat transport depend on low-frequency phonons, while for other properties, such as the free energy, especially high-frequency phonons count. For all these quantities one needs to know the entire phonon band structure, whose simulation becomes exceedingly expensive for more complex systems when using methods like dispersion-corrected density functional theory (DFT). Therefore, in the present contribution we evaluate the performance of more approximate methodologies, including density functional tight binding (DFTB) and a pool of force fields (FF) of varying complexity and sophistication. Beyond merely comparing phonon band structures, we also critically evaluate to what extent derived quantities, like temperature-dependent heat capacities, mean squared thermal displacements and temperature-dependent free energies are impacted by shortcomings in the description of the phonon bands. As a benchmark system, we choose (deuterated) naphthalene, as the only organic semiconductor material for which to date experimental phonon band structures are available in the literature. Overall, the best performance amongst the approximate methodologies is observed for a system-specifically parametrized second-generation force field. Interestingly, in the low-frequency regime also force fields with a rather simplistic model for the bonding interactions (like the General Amber Force Field) perform rather well. As far as the tested DFTB parametrization is concerned, we obtain a significant underestimation of the unit cell volume resulting in a pronounced overestimation of the phonon energies in the low frequency region. This cannot be mended by relying on the DFT-calculated unit cell, since with this unit cell the DFTB phonon frequencies significantly underestimate the experiments.
△ Less
Submitted 6 April, 2020; v1 submitted 7 February, 2020;
originally announced February 2020.
