Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “machine learning tools”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,243 records · Page 69
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Characterizing Machine Learning I/O Workloads on Leadership Scale HPC Systems

High performance computing (HPC) is no longer solely limited to traditional workloads such as simulation and modeling. With the increase in the popularity of machine learning (ML) and deep learning (DL) technologies, we are observing that an increasing number of HPC users are incorporating ML methods into their workflow and scientific discovery processes, across a wide spectrum of science domains such as biology, earth science, and physics. This gives rise to a diverse set of I/O patterns than the traditional checkpoint/restart-based HPC I/O behavior. The details of the I/O characteristics of such ML I/O workloads have not been studied extensively for large-scale leadership HPC systems. This paper aims to fill that gap by providing an in-depth analysis to gain an understanding of the I/O behavior of ML I/O workloads using darshan - an I/O characterization tool designed for lightweight tracing and profiling. We study the darshan logs of more than 23, 000 HPC ML I/O jobs over a time period of one year running on Summit - the second-fastest supercomputer in the world. This paper provides a systematic I/O characterization of ML I/O jobs running on a leadership scale supercomputer to understand how the I/O behavior differs across science domains and the scale of workloads, and analyze the usage of parallel file system and burst buffer by ML I/O workloads.

Paul, Arnab↗
DOE OSTI2025

Deploying and Tracking Software with NCCS Software Provisioning

The National Center for Computational Sciences (NCCS) at Oak Ridge National Laboratory has a long history of deploying ground-breaking leadership-class supercomputers for the U.S. Department of Energy. The latest in this line of supercomputers is Frontier, the first supercomputer to break the exascale barrier (1018 floating-point operations per second) on the TOP500 list. Frontier serves a wide array of scientific domains, from traditional simulation-based workloads to newer AI and Machine Learning workloads. To best serve the NCCS user community, NCCS uses Spack to deploy a comprehensive software stack of scientific software packages, providing straightforward access to these packages through Lmod Environment Modules. Maintaining a large software stack while also including multiple new compiler releases each year is a very time-consuming task. Additionally, it is not straightforward to provide a software stack alongside existing vendor-provided software such as the HPE/Cray Programming Environment (CPE), and existing CPE, Spack, and Lmod integration does not allow for multiple versions of GPU libraries such as AMD’s ROCm to be used. To address these challenges and shortcomings, NCCS has developed the NCCS Software Provisioning tool (NSP)1, a tool for deploying and monitoring software stacks on HPC systems. NSP allows NCCS to quickly and effectively provision software stacks from the ground up using template-driven recipes and configuration files. NSP is successfully deployed on Frontier and several other NCCS clusters, enabling the NCCS software team to quickly deploy software stacks for newly-released compilers, expand current software offerings, better support GPU-based software, and monitor Lmod module usage to identify unused software packages that can be removed from the software stack. In this work, we discuss the shortcomings of the previous CPE, Spack, and Lmod usage at NCCS, provide further details on the implementation and structure of NSP, then discuss the benefits that NSP provides.

Rentschler, Asa [ORNL] (ORCID:0009000597694743)↗
DOE OSTI2023

Status Report on Regulatory Criteria Applicable to the Use of Artificial Intelligence (AI) and Machine Learning (ML)

Although the interest in the use of artificial intelligence (AI) and machine learning (ML) in nuclear energy is increasing rapidly, at present their implementation is limited. This rapid increase in interest is not surprising considering that implementing AI and ML technology would allow for continuous monitoring, facilitate the implementation of predictive maintenance with optimized staffing plans, enable automation and autonomy opportunities that could drastically reduce fixed operation and maintenance costs, and provide training for operations and maintenance. Other industries are using AI for construction, and in the nuclear arena AI could provide great benefit in decommissioning activities. The ability of AI and ML to operate in real time vastly increases their potential impact. Before AI can be used in design, operations, or as a regulatory tool, the specifics on the regulations applicable to the use of AI for nuclear power applications need to be established. The difficulty is that the specific use cases will dictate the applicability of regulations. For example, even within the application domain associated with operations, the regulations might vary if the AI is used to create a virtual reference for plant operations or is used for training, optimization of maintenance intervals, prioritization of maintenance activities, etc. Different still is if the AI is to be used for design or setting technical specifications, which will introduce additional requirements. US Nuclear Regulatory Commission (NRC) licensing reviews are based on an applicant’s design meeting its performance assessment based on (1) safety goals and objectives, (2) deterministic and/or probabilistic analysis of accident scenarios, and (3) quantitative assessment of design alternatives against the safety goals and objectives using accepted engineering tools, methodologies, and performance criteria. The current regulatory framework does not explicitly address AI or autonomous control. However, as implementing AI technology will require the use of a digital platform, it must meet the requirements of an instrumentation and control (I&C) system. The regulatory requirements for AI, which will be incorporated into the I&C system, will be very dependent on how it is used (i.e., its functionality, safety classification, etc.). The licensing process is primarily risk-based with the identification of components and systems as nonsafety, important to safety, or safety related. A risk-informed approach allows further gradation of components and systems based on risk metrics such as core damage frequency or large early release fractions. Thus, the use cases and the risk categorization of impacted systems and components will determine the regulatory requirements. Regardless of how AI is used it presents new opportunities for risk-informing operating, maintenance, and regulatory decisions. Trustworthiness, transparency, and the ability to validate and verify the results will be paramount in showing that the systems and plant still meet their performance requirements. This report describes the results of research to identify regulatory implications of AI technologies and their uses. Specifically, this report reviews current regulatory guidance relevant to the application of AI for design (including design changes or new designs including advanced reactors), construction, operations, training, maintenance, research, testing, and as a regulatory tool. AI can be automated at different levels from purely informative purposes to autonomous controls. The focus of this review included determination of constraints on the application of AI technology, identification of any regulatory gaps or uncertainties, and clarification of anticipated technical basis information likely to be important for regulatory acceptance of these technologies. Currently, any use of AI at nuclear power plants is focused on nonsafety-related applications. The NRC and other regulatory bodies are evaluating providing guidance to address gaps rather than create new regulations to address the use of AI and ML. This approach seems to be the best to encourage AI development without adding regulatory uncertainty.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

REIMAGINING HEAT EXCHANGERS FOR NEXT GENERATION ENVIRONMENTAL SYSTEMS

Air-to-refrigerant heat exchangers (HXs) are essential components in space conditioning, refrigeration, and power systems, and recent efforts have focused on making these devices more compact, reducing refrigerant charge and lowering manufacturing costs. Historically, HX innovation has been limited by available computational resources, design tools, and manufacturing constraints. The best available technologies utilize tube-fin and micro- or macro-channel tubes with fins, which are not necessarily the optimal designs achievable with current technology. In this paper, we highlight the latest advancements in air-to-refrigerant HXs, specifically emphasizing innovations achieved through shape and topology optimization. A multi-scale design optimization approach is introduced, alongside similar methods in literature, which enable highly sophisticated shape-optimized tube designs with more than 50% reduction in size and 25% reduction in refrigerant charge, essential for A3 refrigerant charge limit compliance. The frameworks integrate traditional heat and mass transfer science with state-of-the-art machine learning, genetic algorithms, and adjoint algorithms to create novel designs. While many of these innovative designs may not be manufacturable using conventional methods, they allow us explore the boundaries of what is possible. These novel air-to-refrigerant HXs are key enablers for ultra-low-refrigerant charge heat pump and refrigeration systems.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Updates to the Alliance of Genome Resources central infrastructure

The Alliance of Genome Resources (Alliance) is an extensible coalition of knowledgebases focused on the genetics and genomics of intensively studied model organisms. The Alliance is organized as individual knowledge centers with strong connections to their research communities and a centralized software infrastructure, discussed here. Model organisms currently represented in the Alliance are budding yeast, Caenorhabditis elegans, Drosophila, zebrafish, frog, laboratory mouse, laboratory rat, and the Gene Ontology Consortium. The project is in a rapid development phase to harmonize knowledge, store it, analyze it, and present it to the community through a web portal, direct downloads, and application programming interfaces (APIs). Here, we focus on developments over the last 2 years. Specifically, we added and enhanced tools for browsing the genome (JBrowse), downloading sequences, mining complex data (AllianceMine), visualizing pathways, full-text searching of the literature (Textpresso), and sequence similarity searching (SequenceServer). We enhanced existing interactive data tables and added an interactive table of paralogs to complement our representation of orthology. To support individual model organism communities, we implemented species-specific “landing pages” and will add disease-specific portals soon; in addition, we support a common community forum implemented in Discourse software. We describe our progress toward a central persistent database to support curation, the data modeling that underpins harmonization, and progress toward a state-of-the-art literature curation system with integrated artificial intelligence and machine learning (AI/ML).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

A deep learning interatomic potential developed for atomistic simulation of carbon materials

Interatomic potentials based on neural-network machine learning method have attracted considerable attention in recent years owing to their outstanding ability to balance the accuracy and efficiency in atomistic simulations. In this work, a neural-network potential (NNP) for carbon is generated to simulate the structural properties of various carbon structures. The potential is trained using a database consisting of crystalline and liquid structures obtained by the first-principles density functional theory (DFT) calculations. The developed potential accurately predicts the energies and forces in crystalline and liquid carbon structures, the energetic stability of defected graphene, and the structures of amorphous carbon as the function of density. As a result, the excellent accuracy and transferability of the NNP provide a promising tool for accurate atomistic simulations of various carbon materials with faster speed and much lower cost.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Large Language Models for the Creation and Use of Semantic Ontologies in Buildings: Requirements and Challenges

Semantic ontologies offer a formalized, machine-readable framework for representing knowledge, enabling the structured description of complex systems. In the building domain, the adoption of ontologies like the Brick schema has transformed how buildings and their systems are modeled by providing a standardized, interoperable language. However, the complexity and the steep learning curve involved in developing and querying semantic models present substantial challenges, often requiring a workforce with specialized expertise. This paper builds on our experience in investigating how Large Language Models (LLMs) can help address these challenges, focusing on their role in constructing and querying of semantic models, particularly using the Brick Schema. Our study outlines the requirements and metrics for evaluating the scalability and effectiveness of LLM-based tools, while also discussing the current challenges and limitations in developing such tools. Ultimately, this paper aims to orient research efforts as various groups experiment with diverse techniques, while enabling more effective comparison of emerging solutions and fostering collaboration across the field.

Mulayim, Ozan Baris↗
DOE OSTI2020

Probing Sulfur Chemical and Electronic Structure with Experimental Observation and Quantitative Theoretical Prediction of Kα and Valence-to-Core Kβ X-ray Emission Spectroscopy

An extensive experimental and theoretical study of the Kα and Kβ high-resolution X-ray emission spectroscopy (XES) of sulfur-bearing systems is presented here. This study encompasses a wide range of organic and inorganic compounds, including numerous experimental spectra from both prior published work and new measurements. Employing a linear-response time-dependent density functional theory (LR-TDDFT) approach, strong quantitative agreement is found in the calculation of energy shifts of the core-to-core Kα as well as the full range of spectral features in the valence-to-core Kβ spectrum. The ability to accurately calculate the sulfur Kα energy shift supports the use of sulfur Kα XES as a bulk-sensitive tool for assessing sulfur speciation. The fine structure of the sulfur Kβ spectrum, in conjunction with the theoretical results, is shown to be sensitive to the local electronic structure including effects of symmetry, ligand type and number, and, in the case of organosulfur compounds, to the nature of the bonded organic moiety. This agreement between theory and experiment, augmented by the potential for high-access XES measurements with the latest generation of laboratory-based spectrometers, demonstrates the possibility of broad analytical use of XES for sulfur and nearby third-row elements. The effective solution of the forward problem, i.e., successful prediction of detailed spectra from known molecular structure, also suggests future use of supervised machine learning approaches to experimental inference, as has seen recent interest for interpretation of X-ray absorption near-edge structure (XANES).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Geospatial mapping of distribution grid with machine learning and publicly-accessible multi-modal data

Abstract Detailed and location-aware distribution grid information is a prerequisite for various power system applications such as renewable energy integration, wildfire risk assessment, and infrastructure planning. However, a generalizable and scalable approach to obtain such information is still lacking. In this work, we develop a machine-learning-based framework to map both overhead and underground distribution grids using widely-available multi-modal data including street view images, road networks, and building maps. Benchmarked against the utility-owned distribution grid map in California, our framework achieves > 80% precision and recall on average in the geospatial mapping of grids. The framework developed with the California data can be transferred to Sub-Saharan Africa and maintain the same level of precision without fine-tuning, demonstrating its generalizability. Furthermore, our framework achieves a R 2 of 0.63 in measuring the fraction of underground power lines at the aggregate level for estimating grid exposure to wildfires. We offer the framework as an open tool for mapping and analyzing distribution grids solely based on publicly-accessible data to support the construction and maintenance of reliable and clean energy systems around the world.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Uncovering obscured phonon dynamics from powder inelastic neutron scattering using machine learning

The study of phonon dynamics is pivotal for understanding material properties, yet it faces challenges due to the irreversible information loss inherent in powder inelastic neutron scattering spectra and the limitations of traditional analysis methods. In this study, we present a machine learning framework designed to reveal obscured phonon dynamics from powder spectra. Using a variational autoencoder, we obtain a disentangled latent representation of spectra and successfully extract force constants for reconstructing phonon dispersions. Notably, our model demonstrates effective applicability to experimental data even when trained exclusively on physics-based simulations. The fine-tuning with experimental spectra further mitigates issues arising from domain shift. Analysis of latent space underscores the model’s versatility and generalizability, affirming its suitability for complex system applications. Furthermore, our framework’s two-stage design is promising for developing a universal pre-trained feature extractor. This approach has the potential to revolutionize neutron measurements of phonon dynamics, offering researchers a potent tool to decipher intricate spectra and gain valuable insights into the intrinsic physics of materials.

domain adaptation↗
DOE OSTI2024

Adaptive Methods for Radial Basis Functions

Radial basis functions (RBFs) are a powerful tool for constructing high-order accurate reduced representations of scattered data in arbitrary dimension and on manifolds. We present a method of constructing data approximations in which we utilize a functional tail to capture a global background profile and a RBF neural network (NN) to capture the smaller-scale features. In the RBF NN the RBF centers, matrix shape parameters were selected adaptively for each RBF. We also utilized a geodesic notion of distance on the manifold on which the data lies, e.g., the spherical geodesic for data on the sphere. Although each of these ideas have been been investigated separately in previous works, their combination into a single algorithm is novel. We defined a machine learning problem in which these properties are learned to minimize the data reduction error. We demonstrate the algorithm for applications of scattered data reduction in the plane and on the sphere.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Novel CHI3L1 ‐Associated Angiogenic Phenotypes Define Glioma Microenvironments: Insights From Multi‐Omics Integration

ABSTRACT The CHI3L1 signaling pathway significantly influences glioma angiogenesis, but its role in the tumor microenvironment (TME) remains elusive. We propose a novelCHI3L1‐associated vascular phenotype classification for glioma through integrative analyses of multiple datasets with bulk and single‐cell transcriptome, genomics, digital pathology, and clinical data. We investigated the biological characteristics, genomic alterations, therapeutic vulnerabilities, and immune profiles within these phenotypes through a comprehensive multi‐omics approach. We constructed the vascular‐related risk (VR) score based onCHI3L1‐associated vascular signatures (CAVS) identified by machine learning algorithms. Utilizing unsupervised consensus clustering, gliomas were stratified into three distinct vascular phenotypes: Cluster A, marked by high vascularization and stromal activation with a relatively low levels of tumor‐infiltrating lymphocytes (TILs); Cluster B, characterized by moderate vascularization and stromal activity, coupled with a high density of TILs; and Cluster C, defined by low vascularization and sparse immune cell infiltration. We observed that the CAVS effectively indicated glioma‐associated angiogenesis and immune suppression by single‐cell RNA‐seq analysis. Moreover, the high‐VR‐score group exhibited enhanced angiogenic activity, reduced immune response, resistance to immunotherapy, and poorer clinical outcomes. The VR score independently predicted glioma prognosis and, combined with a nomogram, provided a robust clinical decision‐making tool. Potential drug prediction based on transcription factors for high‐risk patients was also performed. Our study reveals thatCHI3L1‐associated vascular phenotypes shape distinct immune landscapes in gliomas, offering insights for optimizing therapeutic strategies to improve patient outcomes.

Oncology↗
DOE OSTI2023

Rapid Spaceborne Mapping of Wildfire Retardant Drops for Active Wildfire Management

Aerial application of fire retardant is a critical tool for managing wildland fire spread. Retardant applications are carefully planned to maximize fire line effectiveness, improve firefighter safety, protect high-value resources and assets, and limit environmental impact. However, topography, wind, visibility, and aircraft orientation can lead to differences between planned drop locations and the actual placement of the retardant. Information on the precise placement and areal extent of the dropped retardant can provide wildland fire managers with key information to (1) adaptively manage event resources, (2) assess the effectiveness of retardant slowing or stopping fire spread, (3) document location in relation to ecologically sensitive areas; and perform or validate cost-accounting for drop services. This study uses Sentinel-2 satellite data and commonly used machine learning classifiers to test an automated approach for detecting and mapping retardant application. We show that a multiclass model (retardant, burned, unburned, and cloud artifact classes) outperforms a single-class retardant model and that image differencing (post-application minus pre-application) outperforms single-image models. Compared to the random forest and support vector machine, the gradient boosting model performed the best with an overall accuracy of 0.88 and an F1 Score of 0.76 for fire retardant, though results were comparable for all three models. Our approach maps the full areal extent of the dropped retardant within minutes of image availability, rather than linear representations currently mapped by aerial GPS surveys. The development of this capability allows for the rapid assessment of retardant effectiveness and documentation of placement in relation to sensitive environments.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Identification of mobile genetic elements with geNomad

Identifying and characterizing mobile genetic elements in sequencing data is essential for understanding their diversity, ecology, biotechnological applications and impact on public health. Here we introduce geNomad, a classification and annotation framework that combines information from gene content and a deep neural network to identify sequences of plasmids and viruses. geNomad uses a dataset of more than 200,000 marker protein profiles to provide functional gene annotation and taxonomic assignment of viral genomes. Using a conditional random field model, geNomad also detects proviruses integrated into host genomes with high precision. In benchmarks, geNomad achieved high classification performance for diverse plasmids and viruses (Matthews correlation coefficient of 77.8% and 95.3%, respectively), substantially outperforming other tools. Leveraging geNomad’s speed and scalability, we processed over 2.7 trillion base pairs of sequencing data, leading to the discovery of millions of viruses and plasmids that are available through the IMG/VR and IMG/PR databases. geNomad is available at https://portal.nersc.gov/genomad.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Enhanced descriptor identification and mechanism understanding for catalytic activity using a data-driven framework: revealing the importance of interactions between elementary steps

We report accurate identification of descriptors for catalytic activities has long been essential to the in-depth understanding of catalysis and recently to set the basis for catalyst screening. However, commonly used methods suffer from low accuracy in predictability. This study reports an enhanced approach to accurately identify the descriptors from a kinetic dataset using a machine learning (ML) surrogate model. CO hydrogenation to methanol over Cu-based catalysts was taken as a case study. Our model captures not only the contribution from individual elementary steps but also the interaction between relevant steps within a reaction network, which was found to be essential for high accuracy. As a result, six effective descriptors are identified, which are accurate enough to ensure the trained gradient boosted regression (GBR) model for good prediction of the methanol turnover frequency (TOF) over metal (M)-doped Cu(111) model surfaces (M = Au, Cu, Pd, Pt, Ni). More importantly, going beyond the purely mathematical ML model, the catalytic role of each identified descriptor can be revealed by using model-agnostic interpretation tools, which enhances the insight into the promoting effect of alloying. The trained GBR model outperforms the conventional derivative-based methods in terms of both the predictability and the mechanism understanding. It opens alternative possibilities toward accurate descriptor-based rational catalyst optimization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Moltensaltpropnet

MoltenSaltPropnet is a physics-informed machine learning framework that aims to predict the thermophysical properties of molten fluoride and chloride salt mixtures, which are crucial for the design and safety of Generation IV molten salt reactors. The code processes data from the Molten-Salt Thermal Properties Database (MSTDB-TP) and the Janz compendium, converting critically evaluated correlations into fast, differentiable surrogate models for density, viscosity, thermal conductivity, and heat capacity across 448 distinct salt systems. The implementation consists of several key components: 1. Data Curation: The code parses and cleans the raw data, normalizing elemental mole fractions and extracting relevant regression coefficients for various thermophysical properties. 2. Feature Engineering: It generates fixed-length numerical descriptors that encapsulate the composition and temperature, incorporating polynomial interaction terms and dimensionality-reduction techniques to optimize model performance. 3. Coefficient Learning: Four different machine learning architectures are employed: a deep residual network (ResNet), a Kolmogorov–Arnold network (KAN), a sparsity-inducing neural network (SNN), and classical regression models. Each model learns to predict coefficients that define the temperature-dependent correlations for the thermophysical properties. 4. Property Reconstruction: The predicted coefficients are used to compute temperature-dependent property values, ensuring positivity and monotonic trends through a composite loss function that enforces physical constraints. 5. User Interface: An open-source web application enables users to filter the database, train task-specific models, and visualize the results, allowing for rapid exploration of candidate salt mixtures. MoltenSaltPropnet bridges the gap between limited experimental data and high-fidelity reactor simulations, providing a powerful tool for researchers in the field of molten salt reactors and advanced nuclear energy systems.

Retamales, Mauricio Eduardo Tano [Idaho National L↗
DOE OSTI2026

Prediction and Experimental Verification of Electrolyte Solvation Structure from an OMol25-Trained Interatomic Potential

A molecular-level understanding of electrolyte solvation structure and ion–ion correlations is critical to developing next-generation battery chemistries. Atomistic simulation capabilities with sufficient accuracy, speed, and transferability to deliver reliable structural insights while avoiding arduous system-specific reparameterization are thus highly desirable. Machine learning interatomic potentials (MLIPs) trained on large, chemically diverse data sets are revolutionizing computational chemistry, enabling molecular dynamics simulations of battery electrolytes with near-DFT accuracy over 10,000× faster than DFT. While previous MLIP training data sets with suitable elemental coverage for electrolytes have been based on inorganic materials, the Open Molecules 2025 (OMol25) data set provides large-scale molecular DFT MLIP training data with broad elemental coverage and specifically samples tens of millions of electrolyte configurations. Here, we integrate computational modeling with experimental validation to systematically assess the ability of large-scale MLIPs pretrained on materials data or on OMol25 to accurately resolve nanoscale structural organization and ion-solvation characteristics in Na-ion battery electrolytes across diverse physicochemical conditions and compositional regimes. We find that the OMol25-trained Universal Model of Atoms (UMA-OMol) predicts experimentally measured densities and X-ray structure factors in substantially better agreement compared to state-of-the-art models trained only on inorganic materials data. Using UMA-OMol, we further analyze systematic trends in solvation structure as a function of cation identity, anion chemistry, salt concentration, and solvent topology. We observe that increasing system temperature amplifies the heterogeneity within the solvation environment, perturbing cation–solvent interactions and promoting the formation of contact ion pairs (CIPs). Moreover, subtle variations in the solvent topology of glyme-based electrolytes cause pronounced changes in ion correlations and solvation structure. The experimental agreement and microscopic insights shown here position OMol25-trained MLIPs as a practical route to predictive, high-throughput electrolyte simulations beyond the limits of classical force fields and direct DFT molecular dynamics, serving as a powerful tool for accelerating the design of next-generation Na-ion battery electrolytes and beyond.

MLIPs↗
DOE OSTI2023

Ensuring thermodynamic consistency with invertible coarse-graining

Coarse-grained models are a core computational tool in theoretical chemistry and biophysics. A judicious choice of a coarse-grained model can yield physical insights by isolating the essential degrees of freedom that dictate the thermodynamic properties of a complex, condensed-phase system. The reduced complexity of the model typically leads to lower computational costs and more efficient sampling compared with atomistic models. Designing “good” coarse-grained models is an art. Generally, the mapping from fine-grained configurations to coarse-grained configurations itself is not optimized in any way; instead, the energy function associated with the mapped configurations is. In this work, we explore the consequences of optimizing the coarse-grained representation alongside its potential energy function. We use a graph machine learning framework to embed atomic configurations into a low-dimensional space to produce efficient representations of the original molecular system. Because the representation we obtain is no longer directly interpretable as a real-space representation of the atomic coordinates, we also introduce an inversion process and an associated thermodynamic consistency relation that allows us to rigorously sample fine-grained configurations conditioned on the coarse-grained sampling. We illustrate that this technique is robust, recovering the first two moments of the distribution of several observables in proteins such as chignolin and alanine dipeptide.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗