Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “computational intelligence”

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 703 records · Page 39
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

The confluence of machine learning and multiscale simulations

Multiscale modeling has a long history of use in structural biology, as computational biologists strive to overcome the time- and length-scale limits of atomistic molecular dynamics. Contemporary machine learning techniques, such as deep learning, have promoted advances in virtually every field of science and engineering and are revitalizing the traditional notions of multiscale modeling. Deep learning has found success in various approaches for distilling information from fine-scale models, such as building surrogate models and guiding the development of coarse-grained potentials. However, perhaps its most powerful use in multiscale modeling is in defining latent spaces that enable efficient exploration of conformational space. In conclusion, this confluence of machine learning and multiscale simulation with modern high-performance computing promises a new era of discovery and innovation in structural biology.

59 BASIC BIOLOGICAL SCIENCES↗
NASA NTRS1993

Evolution of a high-performance storage system based on magnetic tape instrumentation recorders

In order to provide transparent access to data in network computing environments, high performance storage systems are getting smarter as well as faster. Magnetic tape instrumentation recorders contain an increasing amount of intelligence in the form of software and firmware that manages the processes of capturing input signals and data, putting them on media and then reproducing or playing them back. Such intelligence makes them better recorders, ideally suited for applications requiring the high-speed capture and playback of large streams of signals or data. In order to make recorders better storage systems, intelligence is also being added to provide appropriate computer and network interfaces along with services that enable them to interoperate with host computers or network client and server entities. Thus, recorders are evolving into high-performance storage systems that become an integral part of a shared information system. Data tape has embarked on a program with the Caltech sponsored Concurrent Supercomputer Consortium to develop a smart mass storage system. Working within the framework of the emerging IEEE Mass Storage System Reference Model, a high-performance storage system that works with the STX File Server to provide storage services for the Intel Touchstone Delta Supercomputer is being built. Our objective is to provide the required high storage capacity and transfer rate to support grand challenge applications, such as global climate modeling.

Peters, Bruce↗
DOE OSTI2024

Addressing bias in bagging and boosting regression models

As artificial intelligence (AI) becomes widespread, there is increasing attention on investigating bias in machine learning (ML) models. Previous research concentrated on classification problems, with little emphasis on regression models. This paper presents an easy-to-apply and effective methodology for mitigating bias in bagging and boosting regression models, that is also applicable to any model trained through minimizing a differentiable loss function. Our methodology measures bias rigorously and extends the ML model's loss function with a regularization term to penalize high correlations between model errors and protected attributes. We applied our approach to three popular tree-based ensemble models: a random forest model (RF), a gradient-boosted model (GBT), and an extreme gradient boosting model (XGBoost). We implemented our methodology on a case study for predicting road-level traffic volume, where RF, GBT, and XGBoost models were shown to have high accuracy. Despite high accuracy, the ML models were shown to perform poorly on roads in minority-populated areas. Our bias mitigation approach reduced minority-related bias by over 50%.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Explainable artificial intelligence relates perovskite luminescence images to current-voltage metrics

As the demand for low-cost, high-efficiency solar energy technologies grows, metal halide perovskite (MHP) solar cells have emerged as a promising candidate for next-generation photovoltaics due to their high power conversion efficiencies. However, their poor durability and issues with manufacturing consistency remain significant barriers to commercialization. In this work, we develop deep learning models to support materials characterization and provide insight into features and processes influencing performance. The models are trained using transfer learning of a pretrained model to predict relevant current-voltage (IV) metrics based on different combinations of input electroluminescence (EL) and photoluminescence (PL) images of MHP devices. We examine which image types are most informative in accurately predicting different IV metrics. Additionally, we use explainable artificial intelligence (XAI) techniques to provide insights into specific spatial features in the devices that drive differences in performance. We find that stabilized luminescence images (e.g. those collected after biasing the devices for at least 1 min) are better for predicting metrics of open-circuit voltage (by PL) and short-circuit current (by PL with EL), but that predicting fill factor and overall power output may use the time-evolution of EL images. Based on attribution masks generated by integrated gradients for each device performance metric, we further suggest different loss mechanisms associated with categories of large and small spatial defects. Overall, this case study highlights the potential applicability of XAI methodology for streamlining MHP device analysis and accelerating detailed understanding of the relationships between spatial defects and impacts on performance.

14 SOLAR ENERGY↗
DOE OSTI2023

Airfoil Computational Fluid Dynamics - 2k shapes, 25 AoA's, 3 Re numbers

This dataset contains aerodynamic quantities - including flow field values (momentum, energy, and vorticity) and summary values (coefficients of lift, drag, and momentum) - for 1,830 airfoil shapes computed using the HAM2D CFD (computational fluid dynamics) model. The airfoil shapes were designed using the separable shape tensor parameterization that encodes two-dimensional shapes as elements of the Grassmann manifold. This data-driven approach learns two independent spaces of parameter from a collection of sample airfoils. The first captures large-scale, linear perturbations, and the second defines small-scale, higher-order perturbations. For this dataset, we used the G2Aero database of over 19,000 airfoil shapes to learn a parameter space that captured a wide array of shape characteristics. We sampled airfoil designs over both parameter spaces to explore the full range of possible shape variations. The aerodynamic quantities for the generated airfoil were obtained using the HAM2D code, which is a finite-volume Reynolds-averaged Navier-Stokes (RANS) flow solver. We employ a fifth-order WENO scheme for spatial reconstruction with Roe's flux difference scheme for inviscid flux and second-order central differencing for viscous flux. A preconditioned GMRES method is applied for implicit integration. The Spalart-Allmaras 1-eq turbulence model is used for the turbulence closure, and the Medida-Baeder 2-eq transition model is applied to account for the effects of laminar turbulent transition. The airfoil grid is generated with a total of 400 points on the airfoil surface, the initial wall-normal spacing of y+ = 1, and an outer boundary located at 300 chord lengths away from the wall. The CFD simulations are performed at a freestream Mach number of 0.1, for or three different Reynolds' numbers (3M, 6M, and 9M), and for 25 angles of attack from -4 deg. to 20 deg. with 1 degree increments. Across all these various parameters, this dataset includes the results from over 250,000 CFD simulations. The simulations were performed using the Bridges-2 system at the Pittsburgh Supercomputing Center in February 2023 as part of the INTEGRATE project funded by the Advanced Research Projects Agency - Energy, in the U.S. Department of Energy. The data was collected, reformatted, and preprocessed for this OEDI submission in July 2023 under the Foundational AI for Wind Energy project funded by the U.S. Department of Energy Wind Energy Technologies Office. This dataset is intended to serve as a benchmark against which new artificial intelligence (AI) or machine learning (ML) tools may be tested. Baseline AI/ML methods for analyzing this dataset have been implemented, and a link to their repository containing those models has been provided. The .h5 data file structure can be found in the GitHub Repository resource under explore_airfoil_2k_data.ipynb.

2k↗
DOE OSTI2023

Airfoil Computational Fluid Dynamics - 9k shapes, 2 AoA's

This dataset contains aerodynamic quantities - including flow field values (momentum, energy, and vorticity) and summary values (coefficients of lift, drag, and momentum) - for 8,996 airfoil shapes, computed using the HAM2D CFD (computational fluid dynamics) model. The airfoil shapes were designed using the separable shape tensor parameterization that encodes two-dimensional shapes as elements of the Grassmann manifold. This data-driven approach learns two independent spaces of parameter from a collection of sample airfoils. The first captures large-scale, linear perturbations, and the second defines small-scale, higher-order perturbations. For this data, we used the G2Aero database of over 19,000 airfoil shapes to learn a parameter space that captured a wide array of shape characteristics. We fixed the linear deformations to be the mean over the database and sampled new shapes over a four-dimensional parameter space of higher-order perturbation. This sampling approaches allows for isolated analysis of non-linear airfoil shape deformations while holding other aspects (e.g., airfoil thickness) approximately constant. The aerodynamic quantities for the generated airfoil were obtained using the HAM2D code, which is a finite-volume Reynolds-averaged Navier-Stokes (RANS) flow solver. We employ a fifth-order WENO scheme for spatial reconstruction with Roe's flux difference scheme for inviscid flux and second-order central differencing for viscous flux. A preconditioned GMRES method is applied for implicit integration. The Spalart-Allmaras 1-eq turbulence model is used for the turbulence closure, and the Medida-Baeder 2-eq transition model is applied to account for the effects of laminar turbulent transition. The airfoil grid is generated with a total of 400 points on the airfoil surface, the initial wall-normal spacing of y+ = 1, and an outer boundary located at 300 chord lengths away from the wall. The CFD simulations are performed at a freestream Mach number of 0.1, Reynolds number of 9M, and at two angles of attack, 4 deg. and 12 deg. The simulations were performed using the Bridges-2 system at the Pittsburgh Supercomputing Center in February 2023 as part of the INTEGRATE project funded by the Advanced Research Projects Agency - Energy in the U.S. Department of Energy. The data was collected, reformatted, and preprocessed for this OEDI submission in July 2023 under the Foundational AI for Wind Energy project funded by the U.S. Department of Energy Wind Energy Technologies Office. This dataset is intended to serve as a benchmark against which new artificial intelligence (AI) or machine learning (ML) tools may be tested. Baseline AI/ML methods for analyzing this dataset have been implemented, and a link to their repository containing those models has been provided. The .h5 data file structure can be found in the GitHub Repository resource under explore_airfoil_9k_data.ipynb.

9k↗
DOE OSTI2022

AICCA: AI-Driven Cloud Classification Atlas

Clouds play an important role in the Earth’s energy budget, and their behavior is one of the largest uncertainties in future climate projections. Satellite observations should help in understanding cloud responses, but decades and petabytes of multispectral cloud imagery have to date received only limited use. This study describes a new analysis approach that reduces the dimensionality of satellite cloud observations by grouping them via a novel automated, unsupervised cloud classification technique based on a convolutional autoencoder, an artificial intelligence (AI) method good at identifying patterns in spatial data. Our technique combines a rotation-invariant autoencoder and hierarchical agglomerative clustering to generate cloud clusters that capture meaningful distinctions among cloud textures, using only raw multispectral imagery as input. Cloud classes are therefore defined based on spectral properties and spatial textures without reliance on location, time/season, derived physical properties, or pre-designated class definitions. We use this approach to generate a unique new cloud dataset, the AI-driven cloud classification atlas (AICCA), which clusters 22 years of ocean images from the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Aqua and Terra instruments—198 million patches, each roughly 100 km × 100 km (128 × 128 pixels)—into 42 AI-generated cloud classes, a number determined via a newly-developed stability protocol that we use to maximize richness of information while ensuring stable groupings of patches. AICCA thereby translates 801 TB of satellite images into 54.2 GB of class labels and cloud top and optical properties, a reduction by a factor of 15,000. The 42 AICCA classes produce meaningful spatio-temporal and physical distinctions and capture a greater variety of cloud types than do the nine International Satellite Cloud Climatology Project (ISCCP) categories—for example, multiple textures in the stratocumulus decks along the West coasts of North and South America. We conclude that our methodology has explanatory power, capturing regionally unique cloud classes and providing rich but tractable information for global analysis. AICCA delivers the information from multi-spectral images in a compact form, enables data-driven diagnosis of patterns of cloud organization, provides insight into cloud evolution on timescales of hours to decades, and helps democratize climate research by facilitating access to core data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

Brochure for the DOE Office of Science Workshop on Envisioning Frontiers in AI and Computing for Biological Research

In February of 2025 a joint ASCR/BER workshop was held to identify key transformational research directions for understanding biology using artificial intelligence (AI), digital twins and high-performance (HPC) computational methods to facilitate scientific discovery and innovation in support of the Department of Energy mission. AI technologies offer exciting new groundbreaking methods to analyze large volumes of complex biological data, thereby greatly accelerating the ability to understand, predict, and design biological processes for beneficial purposes. In the laboratory, the bridging of AI-enabled automated experimental technologies, HPC and digital twins will provide potent tools for researchers to explore the fundamental nature of biology and harness its inherent metabolic potential for a variety of beneficial purposes. The focus of this workshop was on how high-performance computational methods can impact this objective by exploring digital twins, foundational models, and data-driven approaches with applications to advance automated laboratory experiments, modeling of complex living systems and engineering new functions into plants and microbial systems relevant to DOE mission. Workshop attendees with expertise in plant science, microbiology, mathematics, computer science, and AI assessed the current state of the science, trends, and AI challenges at the interface of plant and microbial systems biology and computational science to identify opportunities for high-impact research. This collaborative effort capitalized on ASCR's advancements in applied mathematics, computer science, and Exascale systems, and BER's expertise in basic genomics-enabled research on DOE relevant plant and microbial systems. The workshop culminated in four key priority research directions to guide future research and development within DOE Office of Science programs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Chemical classification program synthesis using generative artificial intelligence

Accurately classifying chemical structures is essential for cheminformatics and bioinformatics, including tasks such as identifying bioactive compounds of interest, screening molecules for toxicity to humans, finding non-organic compounds with desirable material properties, or organizing large chemical libraries for drug discovery or environmental monitoring. However, manual classification is labor-intensive and difficult to scale to large chemical databases. Existing automated approaches either rely on manually constructed classification rules, or are deep learning methods that lack explainability. This work presents an approach that uses generative artificial intelligence to automatically write chemical classifier programs for classes in the Chemical Entities of Biological Interest (ChEBI) database. These programs can be used for efficient deterministic run-time classification of SMILES structures, with natural language explanations. The programs themselves constitute an explainable computable ontological model of chemical class nomenclature, which we call the ChEBI Chemical Class Program Ontology (C3PO). We validated our approach against the ChEBI database, and compared our results against deep learning models and a naive SMARTS pattern based classifier. C3PO outperforms the naive classifier, but does not reach the performance of state of the art deep learning methods. However, C3PO has a number of strengths that complement deep learning methods, including explainability and reduced data dependence. C3PO can be used alongside deep learning classifiers to provide an explanation of the classification, where both methods agree. The programs can be used as part of the ontology development process, and iteratively refined by expert human curators.

Artificial Intelligence↗
DOE OSTI2025

ReSpike: A Co-Design Framework for Evaluating SNNs on ReRAM-Based Neuromorphic Processors

With Moore’s law approaching its end, traditional von Neumann architectures are struggling to keep up with the exceeding performance and memory requirements of artificial intelligence and machine learning algorithms. Unconventional computing approaches such as neuromorphic computing that leverage spiking neural networks (SNNs) to perform computation are gaining traction and seek the paradigm shift necessary to sustain the increasing demands of modern applications. Novel memory technologies, such as resistive RAM (ReRAM), employ a crossbar architecture that possesses the inherent capability of efficiently computing vector-matrix multiplication—a dominant operation in SNNs. The prospect of naturally mapping SNNs to the crossbar structures provides a unique opportunity for achieving a high-performance, power-efficient neuromorphic system. In this work, we present ReSpike, which is a new framework, behavioral simulator, and architectural design based on ReRAM crossbar architectures, enabling modeling and co-design to achieve efficient execution of SNNs. We drive this co-design forward by quantifying the impact that ReRAM cell nonidealities have on the corresponding accuracy of an SNN application.

Asifuzzaman, Kazi [ORNL] (ORCID:0000000240044791)↗
DOE OSTI2018

Radiology and Enterprise Medical Imaging Extensions (REMIX)

Radiology and Enterprise Medical Imaging Extensions (REMIX) is a platform originally designed to both support the medical imaging-driven clinical and clinical research operational needs of Department of Radiology of The Ohio State University Wexner Medical Center. REMIX accommodates the storage and handling of “big imaging data,” as needed for large multi-disciplinary cancer-focused programs. The evolving REMIX platform contains an array of integrated tools/software packages for the following: (1) server and storage management; (2) image reconstruction; (3) digital pathology; (4) de-identification; (5) business intelligence; (6) texture analysis; and (7) artificial intelligence. These capabilities, along with documentation and guidance, explaining how to interact with a commercial system (e.g., PACS, EHR, commercial database) that currently exists in clinical environments, are to be made freely available.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2021

Revisiting the ODE Method for Recursive Algorithms: Fast Convergence Using Quasi Stochastic Approximation

Several decades ago, Profs. Sean Meyn and Lei Guo were postdoctoral fellows at ANU, where they shared interest in recursive algorithms. It seems fitting to celebrate Lei Guo's 60th birthday with a review of the ODE Method and its recent evolution. The method has been regarded as a technique for algorithm analysis. It is argued that this viewpoint is backwards: The original stochastic approximation method was surely motivated by an ODE, and tools for analysis came much later (based on establishing robustness of Euler approximations). The paper presents a brief survey of recent research in machine learning that shows the power of algorithm design in continuous time, following by careful approximation to obtain a practical recursive algorithm. While these methods are usually presented in a stochastic setting, this is not a prerequisite. In fact, recent theory shows that rates of convergence can be dramatically accelerated by applying techniques inspired by quasi Monte-Carlo. Subject to conditions, the optimal rate of convergence can be obtained by applying the averaging technique of Polyak and Ruppert. The conditions are not universal, but theory suggests alternatives to achieve acceleration. The theory is illustrated with applications to gradient-free optimization, and policy gradient algorithms for reinforcement learning.

learning and adaptive systems in artificial intell↗
DOE OSTI2026

Velocity Extraction Using Complete Time-Domain Waveform Data and Audio Machine Learning

We developed a new machine learning-based tool for extracting information from interferometry measurements: MIDWAZE (Modular Interferometry Direct Waveform AnalyZEr). This paper showcases MIDWAZE’s ability to extract an object’s velocity information from Photonic Doppler Velocimetry (PDV) data at near-human accuracy with little to no human intervention. MIDWAZE can extract velocities roughly 350 times as fast as a human analyst "rushing" to complete their extractions, with similar extraction accuracy. MIDWAZE’s most outstanding feature is that it operates directly in waveform/temporal space, freeing analysis from certain limitations imposed by traditional spectrogram-based approaches and opening the way to "phase aware" PDV analysis. MIDWAZE also has limited ability to discriminate between different solid objects, which we develop as a first step towards automated discrimination of different kinds of objects such as ejecta clouds.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Automated Extraction of Energy Systems Information from Remotely Sensed Data: A Review and Analysis

We report high quality energy systems information is a crucial input to energy systems research, modeling, and decision-making. Unfortunately, actionable information about energy systems is often of limited availability, incomplete, or only accessible for a substantial fee or through a non-disclosure agreement. Recently, remotely sensed data (e.g., satellite imagery, aerial photography) have emerged as a potentially rich source of energy systems information. However, the use of these data is frequently challenged by its sheer volume and complexity, precluding manual analysis. Recent breakthroughs in machine learning have enabled automated and rapid extraction of useful information from remotely sensed data, facilitating large-scale acquisition of critical energy system variables. Here we present a systematic review of the literature on this emerging topic, providing an in-depth survey and review of papers published within the past two decades. We first taxonomize the existing literature into ten major areas, spanning the energy value chain. Within each research area, we distill and critically discuss major features that are relevant to energy researchers, including, for example, key challenges regarding the accessibility and reliability of the methods. We then synthesize our findings to identify limitations and trends in the literature as a whole, and discuss opportunities for innovation. These include the opportunity to extend the methods beyond electricity to broader energy systems and wider geographic areas; and the ability to expand the use of these methods in research and decision making as satellite data become cheaper and easier to access. We also find that there are persistent challenges: limited standardization and rigor of performance assessments; limited sharing of code, which would improve replicability; and a limited consideration of the ethics and privacy of data.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

The high explosives & affected targets (HEAT) dataset

Artificial Intelligence (AI) surrogate models offer a computationally efficient alternative to full-physics simulations, yet no existing datasets are publicly available for training, testing, and validation of machine learning models of the dynamics of high-explosive driven shocks through multiple materials. Shock propagation through materials is a computationally challenging problem because simulations must include material-specific equations of state (EOS) along with descriptions of other physical processes such as plastic deformation, phase change, damage processes, fluid instabilities, and multi-material interactions. Shocks are typically initiated by high-velocity impacts or explosive loading. The latter case necessitates the addition of models of reactive materials to represent high-explosive (HE) detonation. Here, to address the lack of an expansive dataset for multi-material shock propagation in the AI/ML community, we present the High-Explosives and Affected Targets (HEAT) Dataset. HEAT is a physics-rich collection of two-dimensional, cylindrically symmetric, simulations generated using an Eulerian, multi-material, shock-propagation code developed at Los Alamos National Laboratory. The dataset includes two partitions: (1) the expanding shock-cylinder (CYL) simulations, Figs. 1, and (2) the Perturbed Layered Interface (PLI) simulations, Fig. 2. Entries in both partitions consist of time series of arrays of thermodynamic fields (pressure, density, and temperature), kinematic fields (position and velocity), and additional fields that depend on thermodynamic and/or kinematic fields (e.g., material stress). Materials in the CYL partition include solids (aluminium, copper, depleted uranium, stainless steel, tantalum, and a generic polymer), a liquid (water), gases (air, nitrogen), and a generic detonating material (high explosive, HE). The PLI partition spans a highly varying geometry but consists of fixed materials across entries: Copper, aluminium, stainless steel, generic polymer, and generic HE. HEAT captures critical phenomena such as momentum transfer, shock propagation, plastic deformation, and thermal effects, making HEAT a valuable benchmark for development of AI/ML emulation of multi-material shock propagation.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Synthetic Biology of Plants and Microbes for Agriculture, Environment, and Future Applications

Agriculture is under pressure to provide food for a growing population and the feedstock required to drive the bioeconomy. Methods to breed and genetically modify plants are inadequate to keep pace. When engineering crops, traits are painstakingly introduced into plants one-at-a-time, combine unpredictably, and are continuously expressed. Synthetic biology is changing these paradigms with new genome construction tools, computer aided design (CAD), and artificial intelligence (AI). “Smart plants” contain circuits that respond to environmental change, alter morphology, or respond to threats. Further, the plant and associated microbes (fungi, bacteria, archaea) are now being viewed by genetic engineers as a holistic system. Historically, plant health has been enhanced by many natural and laboratory-evolved soil microbes marketed to enhance growth, provide nutrients, or confer pest/stress resistance. Synthetic biology has expanded the number of species that can be engineered, increased the complexity of engineered functions, controlled environmental release, and assembled stable consortia. New CAD tools will manage genetic engineering projects spanning multiple plant genomes (nucleus, chloroplast, mitochondrion) and the thousands of genomes of associated bacteria/fungi. Here, this review covers advanced genetic engineering techniques to drive the next agricultural revolution, as well as push plant engineering into new realms for manufacturing, infrastructure, sensing, and remediation.

Clauer, Phillip [Massachusetts Inst. of Technology↗
DOE OSTI2021

Quantum-Based Molecular Dynamics Simulations Using Tensor Cores

Tensor cores, along with tensor processing units, represent a new form of hardware acceleration specifically designed for deep neural network calculations in artificial intelligence applications. Tensor cores provide extraordinary computational speed and energy efficiency but with the caveat that they were designed for tensor contractions (matrix–matrix multiplications) using only low-precision floating-point operations. Despite this perceived limitation, we demonstrate how tensor cores can be applied with high efficiency to the challenging and numerically sensitive problem of quantum-based Born–Oppenheimer molecular dynamics, which requires highly accurate electronic structure optimizations and conservative force evaluations. The interatomic forces are calculated on-the-fly from an electronic structure that is obtained from a generalized deep neural network, where the computational structure naturally takes advantage of the exceptional processing power of the tensor cores and allows for high performance in excess of 100 Tflops on a single Nvidia A100 GPU. Stable molecular dynamics trajectories are generated using the framework of extended Lagrangian Born–Oppenheimer molecular dynamics, which combines computational efficiency with long-term stability, even when using approximate charge relaxations and force evaluations that are limited in accuracy by the numerically noisy conditions caused by the low-precision tensor core floating-point operations. A canonical ensemble simulation scheme is also presented, where the additional numerical noise in the calculated forces is absorbed into a Langevin-like dynamics.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

CACTUS: Chemistry Agent Connecting Tool Usage to Science

Large language models (LLMs) have shown remarkable potential in various domains but often lack the ability to access and reason over domain-specific knowledge and tools. In this article, we introduce Chemistry Agent Connecting Tool-Usage to Science (CACTUS), an LLM-based agent that integrates existing cheminformatics tools to enable accurate and advanced reasoning and problem-solving in chemistry and molecular discovery. We evaluate the performance of CACTUS using a diverse set of open-source LLMs, including Gemma-7b, Falcon-7b, MPT-7b, Llama3-8b, and Mistral-7b, on a benchmark of thousands of chemistry questions. Our results demonstrate that CACTUS significantly outperforms baseline LLMs, with the Gemma-7b, Mistral-7b, and Llama3-8b models achieving the highest accuracy regardless of the prompting strategy used. Moreover, we explore the impact of domain-specific prompting and hardware configurations on model performance, highlighting the importance of prompt engineering and the potential for deploying smaller models on consumer-grade hardware without a significant loss in accuracy. By combining the cognitive capabilities of open-source LLMs with widely used domain-specific tools provided by RDKit, CACTUS can assist researchers in tasks such as molecular property prediction, similarity searching, and drug-likeness assessment.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗