-
MetaGFN: Exploring Distant Modes with Adapted Metadynamics for Continuous GFlowNets
Authors:
Dominic Phillips,
Flaviu Cipcigan
Abstract:
Generative Flow Networks (GFlowNets) are a class of generative models that sample objects in proportion to a specified reward function through a learned policy. They can be trained either on-policy or off-policy, needing a balance between exploration and exploitation for fast convergence to a target distribution. While exploration strategies for discrete GFlowNets have been studied, exploration in…
▽ More
Generative Flow Networks (GFlowNets) are a class of generative models that sample objects in proportion to a specified reward function through a learned policy. They can be trained either on-policy or off-policy, needing a balance between exploration and exploitation for fast convergence to a target distribution. While exploration strategies for discrete GFlowNets have been studied, exploration in the continuous case remains to be investigated, despite the potential for novel exploration algorithms due to the local connectedness of continuous domains. Here, we introduce Adapted Metadynamics, a variant of metadynamics that can be applied to arbitrary black-box reward functions on continuous domains. We use Adapted Metadynamics as an exploration strategy for continuous GFlowNets. We show several continuous domains where the resulting algorithm, MetaGFN, accelerates convergence to the target distribution and discovers more distant reward modes than previous off-policy exploration strategies used for GFlowNets.
△ Less
Submitted 2 March, 2025; v1 submitted 28 August, 2024;
originally announced August 2024.
-
Symbolic Learning for Material Discovery
Authors:
Daniel Cunnington,
Flaviu Cipcigan,
Rodrigo Neumann Barros Ferreira,
Jonathan Booth
Abstract:
Discovering new materials is essential to solve challenges in climate change, sustainability and healthcare. A typical task in materials discovery is to search for a material in a database which maximises the value of a function. That function is often expensive to evaluate, and can rely upon a simulation or an experiment. Here, we introduce SyMDis, a sample efficient optimisation method based on…
▽ More
Discovering new materials is essential to solve challenges in climate change, sustainability and healthcare. A typical task in materials discovery is to search for a material in a database which maximises the value of a function. That function is often expensive to evaluate, and can rely upon a simulation or an experiment. Here, we introduce SyMDis, a sample efficient optimisation method based on symbolic learning, that discovers near-optimal materials in a large database. SyMDis performs comparably to a state-of-the-art optimiser, whilst learning interpretable rules to aid physical and chemical verification. Furthermore, the rules learned by SyMDis generalise to unseen datasets and return high performing candidates in a zero-shot evaluation, which is difficult to achieve with other approaches.
△ Less
Submitted 30 November, 2023;
originally announced December 2023.
-
Discovery of Novel Reticular Materials for Carbon Dioxide Capture using GFlowNets
Authors:
Flaviu Cipcigan,
Jonathan Booth,
Rodrigo Neumann Barros Ferreira,
Carine Ribeiro dos Santos,
Mathias Steiner
Abstract:
Artificial intelligence holds promise to improve materials discovery. GFlowNets are an emerging deep learning algorithm with many applications in AI-assisted discovery. By using GFlowNets, we generate porous reticular materials, such as metal organic frameworks and covalent organic frameworks, for applications in carbon dioxide capture. We introduce a new Python package (matgfn) to train and sampl…
▽ More
Artificial intelligence holds promise to improve materials discovery. GFlowNets are an emerging deep learning algorithm with many applications in AI-assisted discovery. By using GFlowNets, we generate porous reticular materials, such as metal organic frameworks and covalent organic frameworks, for applications in carbon dioxide capture. We introduce a new Python package (matgfn) to train and sample GFlowNets. We use matgfn to generate the matgfn-rm dataset of novel and diverse reticular materials with gravimetric surface area above 5000 m$^2$/g. We calculate single- and two-component gas adsorption isotherms for the top-100 candidates in matgfn-rm. These candidates are novel compared to the state-of-art ARC-MOF dataset and rank in the 90th percentile in terms of working capacity compared to the CoRE2019 dataset. We discover 15 materials outperforming all materials in CoRE2019.
△ Less
Submitted 16 October, 2023; v1 submitted 11 October, 2023;
originally announced October 2023.
-
Machine Guided Discovery of Novel Carbon Capture Solvents
Authors:
James L. McDonagh,
Benjamin H. Wunsch,
Stamatia Zavitsanou,
Alexander Harrison,
Bruce Elmegreen,
Stacey Gifford,
Theodore van Kessel,
Flaviu Cipcigan
Abstract:
The increasing importance of carbon capture technologies for deployment in remediating CO2 emissions, and thus the necessity to improve capture materials to allow scalability and efficiency, faces the challenge of materials development, which can require substantial costs and time. Machine learning offers a promising method for reducing the time and resource burdens of materials development throug…
▽ More
The increasing importance of carbon capture technologies for deployment in remediating CO2 emissions, and thus the necessity to improve capture materials to allow scalability and efficiency, faces the challenge of materials development, which can require substantial costs and time. Machine learning offers a promising method for reducing the time and resource burdens of materials development through efficient correlation of structure-property relationships to allow down-selection and focusing on promising candidates. Towards demonstrating this, we have developed an end-to-end "discovery cycle" to select new aqueous amines compatible with the commercially viable acid gas scrubbing carbon capture. We combine a simple, rapid laboratory assay for CO2 absorption with a machine learning based molecular fingerprinting model approach. The prediction process shows 60% accuracy against experiment for both material parameters and 80% for a single parameter on an external test set. The discovery cycle determined several promising amines that were verified experimentally, and which had not been applied to carbon capture previously. In the process we have compiled a large, single-source data set for carbon capture amines and produced an open source machine learning tool for the identification of amine molecule candidates (https://github.com/IBM/Carbon-capture-fingerprint-generation).
△ Less
Submitted 24 March, 2023;
originally announced March 2023.
-
Accelerating Antimicrobial Discovery with Controllable Deep Generative Models and Molecular Dynamics
Authors:
Payel Das,
Tom Sercu,
Kahini Wadhawan,
Inkit Padhi,
Sebastian Gehrmann,
Flaviu Cipcigan,
Vijil Chenthamarakshan,
Hendrik Strobelt,
Cicero dos Santos,
Pin-Yu Chen,
Yi Yan Yang,
Jeremy Tan,
James Hedrick,
Jason Crain,
Aleksandra Mojsilovic
Abstract:
De novo therapeutic design is challenged by a vast chemical repertoire and multiple constraints, e.g., high broad-spectrum potency and low toxicity. We propose CLaSS (Controlled Latent attribute Space Sampling) - an efficient computational method for attribute-controlled generation of molecules, which leverages guidance from classifiers trained on an informative latent space of molecules modeled u…
▽ More
De novo therapeutic design is challenged by a vast chemical repertoire and multiple constraints, e.g., high broad-spectrum potency and low toxicity. We propose CLaSS (Controlled Latent attribute Space Sampling) - an efficient computational method for attribute-controlled generation of molecules, which leverages guidance from classifiers trained on an informative latent space of molecules modeled using a deep generative autoencoder. We screen the generated molecules for additional key attributes by using deep learning classifiers in conjunction with novel features derived from atomistic simulations. The proposed approach is demonstrated for designing non-toxic antimicrobial peptides (AMPs) with strong broad-spectrum potency, which are emerging drug candidates for tackling antibiotic resistance. Synthesis and testing of only twenty designed sequences identified two novel and minimalist AMPs with high potency against diverse Gram-positive and Gram-negative pathogens, including one multidrug-resistant and one antibiotic-resistant K. pneumoniae, via membrane pore formation. Both antimicrobials exhibit low in vitro and in vivo toxicity and mitigate the onset of drug resistance. The proposed approach thus presents a viable path for faster and efficient discovery of potent and selective broad-spectrum antimicrobials.
△ Less
Submitted 25 February, 2021; v1 submitted 22 May, 2020;
originally announced May 2020.
-
Electronic coarse graining enhances the predictive power of molecular simulation allowing challenges in water physics to be addressed
Authors:
Flaviu S. Cipcigan,
Vlad P. Sokhan,
Jason Crain,
Glenn J. Martyna
Abstract:
One key factor that limits the predictive power of molecular dynamics simulations is the accuracy and transferability of the input force field. Force fields are challenged by heterogeneous environments, where electronic responses give rise to biologically important forces such as many-body polarisation and dispersion. The importance of polarisation was recognised early-on and described by Cochran…
▽ More
One key factor that limits the predictive power of molecular dynamics simulations is the accuracy and transferability of the input force field. Force fields are challenged by heterogeneous environments, where electronic responses give rise to biologically important forces such as many-body polarisation and dispersion. The importance of polarisation was recognised early-on and described by Cochran in 1959 [Philosophical Magazine 4 (1959) 1082-1086]. However, dispersion forces are still treated at the two-body level and in the dipole limit, although the importance of three-body terms in the condensed phase was demonstrated by Barker in the 1980s [Phys. Rev. Lett. 57 (1986) 230-233]. A way of treating both polarisation and dispersion on an equal basis is to coarse grain the electrons a molecular moiety to a single quantum harmonic oscillator, as suggested as early as the 1960s by Hirschfelder, Curtiss and Bird [The Molecular Theory of Gases and Liquids (1954)]. This treatment, when solved in the strong coupling limit, gives all orders of long-range forces. In the last decade, the tools necessary to exploit this strong coupling limit have been developed, culminating in a transferable model of water with excellent predictive power across the phase diagram. This transferability arises since the environment identifies the form of long range interactions, rather than the expressions selected by the modeller. Here, we discuss the role of electronic coarse-graining in predictive multiscale materials modelling and describe the first implementation of the method in a general purpose molecular dynamics software, QDO_MD.
△ Less
Submitted 10 September, 2016;
originally announced September 2016.
-
Molecular-scale remnants of the liquid-gas transition in supercritical polar fluids
Authors:
V. P. Sokhan,
A. Jones,
F. S. Cipcigan,
J. Crain,
G. J. Martyna
Abstract:
An electronically coarse-grained model for water reveals a persistent vestige of the liquid-gas transition deep into the supercritical region. A crossover in the density dependence of the molecular dipole arises from the onset of non-percolating hydrogen bonds. The crossover points coincide with the Widom line in the scaling region but extend further, tracking the heat capacity maxima, offering ev…
▽ More
An electronically coarse-grained model for water reveals a persistent vestige of the liquid-gas transition deep into the supercritical region. A crossover in the density dependence of the molecular dipole arises from the onset of non-percolating hydrogen bonds. The crossover points coincide with the Widom line in the scaling region but extend further, tracking the heat capacity maxima, offering evidence for liquid- and gas-like state points in a "one-phase" fluid. The effect is present even in dipole-limit models suggesting that it is common for all molecular liquids exhibiting dipole enhancement in the liquid phase.
△ Less
Submitted 5 August, 2015;
originally announced August 2015.
