Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Algorithms and data structure”

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 1,621 records · Page 90

Measuring the Energy Consumption and Efficiency of Deep Neural Networks: An Empirical Analysis and Design Recommendations

Addressing the "Red-AI" trend of rising energy consumption by large-scale neural networks, this study investigates the measured energy consumption of training various fully connected neural network architectures. We introduce the BUTTER-E dataset, an augmentation to the BUTTER Empirical Deep Learning dataset, containing energy consumption and performance data from 41,129 individual experimental runs spanning 30,582 distinct configurations: 13 datasets, 20 sizes (trainable parameters), 8 "shapes", and 14 depths on both CPUs and GPUs using node-level watt-meters. This dataset reveals the complex relationship between dataset size, network structure, and energy use. Our analysis uncovers a surprising, hardware-mediated non-linear relationship between energy efficiency and network design, challenging the assumption that reducing the number of parameters or FLOPs is the best way to achieve greater energy efficiency. We propose a straightforward and effective energy model that accounts for network size, computing, and memory hierarchy. Highlighting the need for cache-considerate algorithm development, we suggest a codesign approach to energy efficient network, algorithm, and hardware design. This work contributes to the fields of sustainable computing and Green AI, offering practical guidance for creating more energy-efficient neural networks and promoting sustainable AI.

97 MATHEMATICS AND COMPUTING↗

Identifying atmospheric rivers and their poleward latent heat transport with generalizable neural networks: ARCNNv1

Abstract. Atmospheric rivers (ARs) are extreme weather events that can alleviate drought or cause billions of US dollars in flood damage. By transporting significant amounts of latent energy towards the poles, they are crucial to maintaining the climate system's energy balance. Since there is no first-principle definition of an AR grounded in geophysical fluid mechanics, AR identification is currently performed by a multitude of expert-defined, threshold-based algorithms. The variety of AR detection algorithms has introduced uncertainty into the study of ARs, and the thresholds of the algorithms may not generalize to new climate datasets and resolutions. We train convolutional neural networks (CNNs) to detect ARs while representing this uncertainty; we name these models ARCNNs. To detect ARs without requiring new labeled data and labor-intensive AR detection campaigns, we present a semi-supervised learning framework based on image style transfer. This framework generalizes ARCNNs across climate datasets and input fields. Using idealized and realistic numerical models, together with observations, we assess the performance of the ARCNNs. We test the ARCNNs in an idealized simulation of a shallow-water fluid in which nearly all the tracer transport can be attributed to AR-like filamentary structures. In reanalysis and a high-resolution climate model, we use ARCNNs to calculate the contribution of ARs to meridional latent heat transport, and we demonstrate that this quantity varies considerably due to AR detection uncertainty.

54 ENVIRONMENTAL SCIENCES↗

Development of Acoustic Model-Based Iterative Reconstruction Technique for Thick-Concrete Imaging

Ultrasound signals have been used extensively for non-destructive evaluation (NDE). However, typical reconstruction techniques, such as the synthetic aperture focusing technique (SAFT), are limited to quasi-homogenous thin media. New ultrasonic systems and reconstruction algorithms are in need for one-sided NDE of non-homogenous thick objects. An application example space is imaging of reinforced concrete structures for commercial nuclear power plants (NPPs). These structures provide important foundation, support, shielding, and containment functions. Identification and management of aging and degradation of concrete structures is fundamental to the proposed long-term operation of NPPs. Another example is geothermal and oil/gas production wells. These multi-layered structures are composed of steel, cement, and several types of soil and rocks. Ultrasound systems with greater penetration range and image quality will allow for better monitoring of the well s health and prediction of high-pressure hydraulic fracturing of the rock. These application challenges need to be addressed with an integrated imaging approach, where the application, hardware, and reconstruction software are highly integrated and optimized. Therefore, we are developing an ultrasonic system with Model-Based Iterative Reconstruction (MBIR) as the image reconstruction backbone. As the first implementation of MBIR for ultrasonic signals, this paper document the first implementation of the algorithm and show reconstruction results for synthetically generated data.

Almansouri, Hani↗

Implementation of a Novel Algorithm For Generating Synthetic CT Images From Magnetic Resonance Imaging Data Sets for Prostate Cancer Radiation Therapy

Purpose: To describe and evaluate a method for generating synthetic computed tomography (synCT) images from magnetic resonance simulation (MR-SIM) data for accurate digitally reconstructed radiograph (DRR) generation and dose calculations in prostate cancer radiation therapy. Methods and Materials: A retrospective evaluation was performed in 9 prostate cancer patients who had undergone MR-SIM in addition to CT simulation (CT-SIM). MR-SIM data were used to generate synCT images by using a novel, voxel-based weighted summation approach. A subset of patients was used for weight optimization, and the number of patients to use during optimization was determined. Hounsfield unit (HU) differences between CT-SIM and synCT images were analyzed via mean absolute error (MAE). Original, CT-based treatment plans were mapped onto synCTs. DRRs were generated, and agreement between CT and synCT-generated DRRs was evaluated via Dice similarity coefficient (DSC). Dose was recalculated, and dose-volume metrics and gamma analysis were used to evaluate resulting treatment plans. Results: Full field-of-view synCT MAE across all patients was 74.3 ± 10.9 HU with differences from CTs of 2.0 ± 8.1 HU and 11.9 ± 46.7 HU for soft tissue structures (prostate, bladder, and rectum) and femoral bones, respectively. Calculated DSCs for anterior-posterior and lateral DRRs were 0.90 ± 0.04 and 0.92 ± 0.05, respectively. Differences in D99%, mean dose, and maximum dose to the clinical target volume from CT-SIM dose calculations were 0.75% ± 0.35%, 0.63% ± 0.34%, and 0.54% ± 0.33%, respectively, for synCT-generated plans. Gamma analysis (2%/2 mm dose difference/distance to agreement) revealed pass rates of 99.9% ± 0.1% (range, 99.7%-100%). Conclusion: Generated synCTs enabled accurate DRR generation and dose computation for prostate MR-only simulation. Dose recalculated on synCTs agreed well with original planning distributions. Further validation using a larger patient cohort is warranted.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Fallon FORGE: Distinct Element Reservoir Modeling

Archive containing input/output data for distinct element reservoir modeling for Fallon FORGE. Models created using 3DEC, InSite, and in-house Python algorithms (ITASCA). List of archived files follows; please see 'Modeling Metadata.pdf' (included as a resource below) for additional file descriptions. Data sources include regional geochemical model, well positions and geometry, principal stress field, capability for hydraulic fractures, capability for hydro-shearing, reservoir geomechanical model-stimulation into multiple zones, modeled thermal behavior during circulation, and microseismicity.

15 GEOTHERMAL ENERGY↗

Atmospheric Turbulent Intermittency Over the Arctic Sea‐Ice Surface During the MOSAiC Expedition

Turbulent motions in the Arctic stable boundary layer are characterized by intermittency, but they are rarely investigated due to limited observations, in particular over the sea‐ice surface. In the present study, we explore the characteristics of turbulent intermittency over the Arctic sea‐ice surface using data collected during the Multidisciplinary drifting Observation for the Study of Arctic Climate expedition from October 2019 to September 2020. We first develop a new algorithm, which performs well in identifying the spectral gap over the Arctic sea‐ice surface. Then the characteristics of intermittency are investigated. It is found that the strength of intermittency increases under the conditions of light surface wind speed, small surface wind speed gradient, and strong surface air temperature gradient. The momentum flux, sensible heat flux, and latent heat flux calculated by raw eddy‐covariance fluctuations are overestimated by 3%, 10%, and 24%, respectively, because submesoscale motions are included. Furthermore, the characteristics of the atmospheric boundary layer structure under various intermittency conditions reveal that strong low‐level jets are favorable to surface turbulent motions that result in weak intermittency, while strong temperature inversions above the surface layer suppress surface turbulent motions and lead to strong intermittency.

54 ENVIRONMENTAL SCIENCES↗

The big data challenges of connectomics

The structure of the nervous system is extraordinarily complicated because individual neurons are interconnected to hundreds or even thousands of other cells in networks that can extend over large volumes. Mapping such networks at the level of synaptic connections, a field called connectomics, began in the 1970s with a the study of the small nervous system of a worm and has recently garnered general interest thanks to technical and computational advances that automate the collection of electron-microscopy data and offer the possibility of mapping even large mammalian brains. However, modern connectomics produces ‘big data’, unprecedented quantities of digital information at unprecedented rates, and will require, as with genomics at the time, breakthrough algorithmic and computational solutions. Here in this paper we describe some of the key difficulties that may arise and provide suggestions for managing them.

60 APPLIED LIFE SCIENCES↗

Structural evolution of two-dimensional silicates using a “bond-switching” algorithm

Silicates are the most abundant materials in the earth's crust. In recent years, two-dimensional (2D) versions of them grown on metal supports (known as bilayer silicates) have allowed their study in detail down to the atomic scale. These structures are self-containing. Furthermore, they are not covalently bound to the metal support but interact with it through van der Waals forces. Like their three-dimensional counterparts, the 2D-silicates can form both crystalline and vitreous structures. Furthermore, the interconversion between vitreous to crystalline structures has been experimentally observed at the nanoscale. While theoretical work has been carried out to try to understand these transformations, a limitation for ab initio methods, and even molecular dynamics methods, is the computational cost of studying large systems and long timescales. In this work, we present a simple and computationally inexpensive approach, that can be used to represent the evolution of bilayer silicates using a bond-switching algorithm. This approach allows reaching equilibrium ring size distributions as a function of a parameter that can be related to the ratio between temperature and the energy required for the bond-switching event. The ring size distributions are compared to experimental data available in the literature.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

Ultrafast processing of pixel detector data with machine learning frameworks

Modern photon science performed at high repetition rate free-electron laser (FEL) facilities and beyond relies on 2D pixel detectors operating at increasing frequencies (towards 100 kHz at LCLS-II) and producing rapidly increasing amounts of data (towards TB/s). This data must be rapidly stored for offline analysis and summarized in real time for online feedback to the scientists. While at LCLS all raw data has been stored, at LCLS-II this would lead to a prohibitive cost; instead, enabling real time processing of pixel detector data (dark, gain, common mode, background, charge summing, subpixel position, photon counting, data summarization) allows reducing the size and cost of online processing, offline processing and storage by orders of magnitude while preserving full photon information. This could be achieved by taking advantage of the compressibility of sparse data typical for LCLS-II applications. Faced with a similar big data challenge a decade ago, computer vision stimulated revolutionary advances in machine learning hardware and software. We investigated if these developments are useful in the field of data processing for high speed pixel detectors and found that typical deep learning models and autoencoder architectures failed to yield useful noise reduction while preserving full photon information, presumably because of the very different statistics and feature sets in computer vision and radiation imaging. However, the raw performance of modern frameworks like Tensorflow inspired us to redesign in Tensorflow mathematically equivalent versions of the state-of-the-art, “classical” algorithms used at LCLS. The novel Tensorflow models resulted in elegant, compact and hardware agnostic code, gaining 1 to 2 orders of magnitude faster processing on an inexpensive consumer GPU, reducing by 3 orders of magnitude the projected cost of online analysis and compression without photon loss at LCLS-II. The novel Tensorflow models also enabled ongoing development of a pipelined hardware system expected to yield an additional 3 to 4 orders of magnitude speedup, necessary for meeting the data acquisition and storage requirements at LCLS-II, potentially enabling acquiring every single FEL pulse at full speed. Computer vision a decade ago was dominated by hand-crafted filters; their structure inspired the deep learning revolution resulting in modern deep convolutional networks; similarly, our novel Tensorflow filters provide inspiration for designing future deep learning models for ultrafast and efficient processing and classification of pixel detector images at FEL facilities.

47 OTHER INSTRUMENTATION↗

A science-driven approach to optimize the design for a biological small-angle neutron scattering instrument

Biological small-angle neutron scattering (SANS) instruments facilitate critical analysis of the structure and dynamics of complex biological systems. However, with the growth of experimental demands and the advances in optical systems design, a new neutron optical concept is necessary to overcome the limitations of current instruments. This work presents an approach to include experimental objectives ( i.e. the science to be supported by a specific neutron scattering instrument) in the optimization of the neutron optical concept. The approach for a proposed SANS instrument at the Second Target Station of the Spallation Neutron Source at Oak Ridge National Laboratory, USA, is presented here. Further, the instrument is simulated with the McStas software package. The optimization process is driven by an evolutionary algorithm using McStas output data, which are processed to calculate an objective function designed to quantify the expected performance of the simulated neutron optical configuration for the intended purpose. Each McStas simulation covers the complete instrument, from source to detector, including realistic sample scattering functions. This approach effectively navigates a high-dimensional parameter space that is otherwise intractable; it allows the design of next-generation SANS instruments to address specific scientific cases and has the potential to increase instrument performance compared with traditional design approaches.

47 OTHER INSTRUMENTATION↗

Modeling and Optimization of Zeolites for Contaminant Removal from Coal Combustion Impoundment Leachates

Coal Combustion Residual (CCR) landfills in the U.S. can lead to elevated contaminant concentrations in groundwater and runoff, particularly of arsenic, boron, and selenium. Zeolites can be readily synthesized from materials contained within the coal combustion ash impoundments and can be cation-exchanged to enhance adsorption capacity, selectivity, and reactivity. However, the optimization of zeolites becomes a daunting task when accounting for the variety of Si:Al ratios, the species of extra-framework cations present, and the zeolite pore structure. Molecular simulations provide methods to study and guide the design of zeolites for the sorption of contaminants from aqueous solution. In this work, models that predict the sorption of arsenic, boron, and selenium from water using cation-exchanged zeolites. Because of the lack of experimental adsorption data for these species of contaminants, models were parametrized to reproduce the results of first-principles calculations and then used to predict sorption for zeolites for a dataset containing 6000 combinations of zeolites and sorption conditions. Machine learning was used to train a model to predict sorption for materials in this database based on the results of the molecular simulations. Next, a genetic algorithm was used to optimize zeolites for the removal of each contaminant from aqueous solution for individual impoundment sites based on contaminant concentrations reported by the Electric Power Research Institute.

Findley, John↗

Machine Learning Prediction of Fracture Toughness in Hydrogen-charged Stainless Steels

Austenitic stainless steels are structural materials utilized in tritium gas pressure boundaries since they are resistant to hydrogen isotope embrittlement [1-3]. However, exposure to tritium over long periods of time leads to tritium uptake which decays to result in helium ingrowth. This helium ingrowth results in further embrittlement effects which are synergistic with that from the hydrogen isotope [4]. Therefore, it is important for tritium facilities to understand the material limitations of stainless steel in this environment. The Savannah River National Laboratory (SRNL) has available a large experimental data set of austenitic stainless steels which have been exposed to tritium environments for various lengths of time. With the availability of this data set, machine learning (ML) algorithms provide an opportunity to model the embrittlement of stainless steel due to the algorithm’s ability to identify patterns in data sets that are difficult and costly to identify in other manners [5]. Ultimately, the amount and quality of the available data is one defining force in the ability of a ML model to accurately predict the desired outputs. The models developed herein will illustrate the ability for the various algorithms to predict the change in fracture toughness in stainless steels due to hydrogen-isotope embrittlement.

Hoar, Eric T.↗

Multi-Source Machine Learning and Thermoplastics Enhanced Aerostructure Manufacturing (mTEAM)

RTX Technology Research Center (RTRC), together with Collins Aerospace (Collins) and Oak Ridge National Laboratory (ORNL) has developed an Artificial Intelligence (AI) / Machine Learning (ML) guided solution to advance the manufacturing and assembly of high performance and lightweight thermoplastic composite (TPC) aerospace products. The solution aims to lower risk, cost and lead time for induction heating based welding and consolidation processes for TPC structure. The cost and lead time of part and material specific process development for induction welding (IW) and induction consolidation will be reduced by replacing traditional empirical methods with optimization methods that merge AI/ML and physics-based process simulations and process experiments with sensing and controls. TPC-IW process development is empirical in nature, and uncertainties in material & process behavior exist near & far from the induction coil. Physics-based simulations can be leveraged directly for process optimization but can be too computationally expensive to run in high fidelity and real time to do robust process optimization. The key impact of successful TPC induction consolidation and welding is cost & lead time reduction for part & material specific consolidation and welding recipes. This is an enabler for more rapid deployment of TPC structures via joining assembly, which can reduce energy & cost intensive usage of autoclaves & ovens. The solution aimed to advance the U.S. Department of Energy’s interests in using thermoplastics and automation in composite manufacturing for improvement of products for existing markets via increased production speeds, reduced costs, and lowered use of energy. Welded TPC structures can offer significant weight & energy savings for high-value commercial aerospace & industrial applications compared to metal & thermoset composite structures assembled by mechanical fastening and/or adhesive bonding. The project was organized into two Budget Periods. Budget Period 1 (BP1) was 15 months and its goal was to perform ML process optimization framework development & deployment on lab-coupon aerostructure components. A Go/No-Go Review was performed at the end of BP1 to verify fulfilment of key tasks & milestones to justify a Go Decision to move into the next Budget Period. Budget Period 2 (BP2) was 12 months and its goal was the deployment of the ML framework for ML process optimization of pilot industrial scale aerostructure components. The overall project aim was to develop & demonstrate ML-enhanced modeling framework that learns process-property mapping from multiple data sources at different fidelities. During BP1, the team accomplished key tasks & milestones to demonstrate the concept of multi-source ML for TPC aerostructure consolidation and assembly. First, the team completed documentation of induction based TPC heating requirements including baseline metrics to compare measured results against. Next the team completed demonstration of data generation from physics-based simulations for ML surrogate model generation and demonstrated the integration of physics-based simulation data into multi-source AI/ML algorithms. In parallel, the team established the lab-coupon scale induction welding system and completed a process to label and reduce generated data from physics-based simulation and experiments for ML surrogate models to enable multi-source ML model training & testing. To complete BP1, the team integrated physics-based simulation data and experimental data into multi-source ML algorithms. This was based on the team completing ML deployment of the induction welding on a lab system at RTRC and AI/ML deployment on existing induction welding line at Collins. ORNL visited both Collins and RTRC sites to witness the TPC induction welding process. Then, ORNL designed and constructed a new version of their vision-based sensing system better adapted to acquire process signals of the TPC induction welding process for process anomaly and defect detection. In BP2, the team accomplished key tasks & milestones to scale up multi-source ML for TPC aerostructure consolidation and assembly from the lab-coupon scale to the pilot-industrial scale. In BP2, the team demonstrated real time anomaly & defect detection via experiments performed by ORNL & RTRC. The team completed ML-optimization heating trials for TPC induction consolidation at Collins, and the team confirmed pilot industrial scale experimental data from Collins was compatible with the developed ML pipeline from RTRC. The team completed sub-element scale ML process optimization demonstration at RTRC, where the team leveraged RTRC’s robotic TPC welding setup to de-risk the ML process optimization by performing ML analysis of recorded temperatures to account for complex part features. Then, the team applied its ML-derived control strategies and ML process optimization framework at Collins to the pilot-industrial scale on a demo skin-stiffener part representative of a nacelle aerostructure fan cowl section. The key innovation is the AI/ML framework enabling effective process development of high performance, lightweight, energy efficient TPCs for composite aircraft structures.

36 MATERIALS SCIENCE↗

Impurity gas detection for SNF canisters using probabilistic deep learning and acoustic sensing *

Abstract Monitoring impurity gases in spent nuclear fuel (SNF) canisters is a novel structural health monitoring approach for SNF in dry storage. The SNF canisters are sealed containers that do not facilitate visual access to the inside. Acoustic sensing can be deployed by taking advantage of the pathways unobstructed by internal hardware. Although the ultrasonic time-of-flight measurement can provide valuable information, it is limited in its ability to discern the concentration of only one impurity gas. As such, deep learning algorithms, particularly convolutional neural networks (CNNs), offer a promising solution. In this study, CNN-based probabilistic deep learning models were implemented to detect and quantify multiple impurity gases in helium. An experimental platform was established to simulate canister conditions, and ultrasonic test data were collected. The presence of argon and air in helium at concentrations ranging from 0% to 1.2% at increments of 0.05% was considered. The multi-layer perceptron, decision tree, and logistic regression classifiers achieved high accuracies when distinguishing pure helium from helium with impurities. CNN with dropout layers and CNN using maximum likelihood estimation showed a similar performance, indicating their ability to capture uncertainties. The ensemble CNN model exhibited improved predictions and the ability to balance individual gas concentration by integrating 1D- and 2D-CNN models. These findings contribute probabilistic deep learning solutions for impurity gas detection and analysis within SNF canisters, thus ensuring safe storage and management of SNFs.

47 OTHER INSTRUMENTATION↗

SU-E-J-182: A Feasibility Study Evaluating Automatic Identification of Gross Tumor Volume for Breast Cancer Radiotherapy Using Dynamic Contrast-Enhanced MR Imaging

Purpose: To develop a computerized pharmacokinetic model-free Gross Tumor Volume (GTV) segmentation method based on dynamic contrastenhanced MRI (DCE-MRI) data that can improve physician GTV contouring efficiency. Methods: 12 patients with biopsy-proven early stage breast cancer with post-contrast enhanced DCE-MRI images were analyzed in this study. A fuzzy c-means (FCM) clustering-based method was applied to segment 3D GTV from pre-operative DCE-MRI data. A region of interest (ROI) is selected by a clinician/physicist, and the normalized signal evolution curves were calculated by dividing the signal intensity enhancement value at each voxel by the pre-contrast signal intensity value at the corresponding voxel. Three semi-quantitative metrics were analyzed based on normalized signal evolution curves: initial Area Under signal evolution Curve (iAUC), Immediate Enhancement Ratio (IER), and Variance of Enhancement Slope (VES). The FCM algorithm wass applied to partition ROI voxels into GTV voxels and non-GTV voxels by using three analyzed metrics. The partition map for the smaller cluster is then generated and binarized with an automatically calculated threshold. To reduce spurious structures resulting from background, a labeling operation was performed to keep the largest three-dimensional connected component as the identified target. Basic morphological operations including hole-filling and spur removal were useutilized to improve the target smoothness. Each segmented GTV was compared to that drawn by experienced radiation oncologists. An agreement index was proposed to quantify the overlap between the GTVs identified using two approaches and a thershold value of 0.4 is regarded as acceptable. Results: The GTVs identified by the proposed method were overlapped with the ones drawn by radiation oncologists in all cases, and in 10 out of 12 cases, the agreement indices were above the threshold of 0.4. Conclusion: The proposed automatic segmentation method was shown to be promising and might be used to improve physician contouring efficiency. J Horton receives grant from NIH and Varian Medical Systems; F-F Yin receives grant from Varian Medical Systems.

60 APPLIED LIFE SCIENCES↗

Determination of the Spectral Index in the Fission Spectrum Energy Regime

Neutron reaction cross sections play a vital role in tracking the production and destruction of isotopes exposed to neutron fluence. They are central to the process of reconciling the initial and final atom inventories. Measurements of irradiated samples by radiochemical methods in tangent with an algorithm are used to evaluate the fluence a sample is exposed to over the course of the irradiation. This algorithm is the Isotope Production Code (IPC) created and used by the radiochemistry data assessment team at Los Alamos National Laboratory (LANL). An integral result is calculated by varying the total neutron fluence seen by a sample. A sample, irradiated in a critical assembly, will be exposed to a unique neutron flux defined by the neutron source and distance of the sample from the source. Neutron cross sections utilized are a function of the hardness of the neutron spectrum at the location of irradiation. A spectral index is used an indicator of the hardness of the neutron spectrum. Cross sections fit forms applied in IPC are collapsed from a LANL 30-group energy structure. Several decades of research and development have been performed to formalize the current IPC cross section library. Basis of the current fission spectrum neutron reaction cross section library is rooted in critical assembly experiments performed from the 1950’s through the early 1970’s at LANL. The focus of this report is development of the spectral index used an indicator of the hardness of the neutron spectrum in the fission spectrum energy regime.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

HDBind: encoding of molecular structure with hyperdimensional binary representations

Traditional methods for identifying “hit” molecules from a large collection of potential drug-like candidates rely on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have a significant limitation in that they require exceptional computing capabilities for even relatively small collections of molecules. Increasingly large and complex state-of-the-art deep learning approaches have gained popularity with the promise to improve the productivity of drug design, notorious for its numerous failures. However, as deep learning models increase in their size and complexity, their acceleration at the hardware level becomes more challenging. Hyperdimensional Computing (HDC) has recently gained attention in the computer hardware community due to its algorithmic simplicity relative to deep learning approaches. The HDC learning paradigm, which represents data with high-dimension binary vectors, allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas (computer vision, bioinformatics, mass spectrometery, remote sensing, edge devices, etc.). To the best of our knowledge, our work is the first to consider HDC for the task of fast and efficient screening of modern drug-like compound libraries. We also propose the first HDC graph-based encoding methods for molecular data, demonstrating consistent and substantial improvement over previous work. We compare our approaches to alternative approaches on the well-studied MoleculeNet dataset and the recently proposed LIT-PCBA dataset derived from high quality PubChem assays. We demonstrate our methods on multiple target hardware platforms, including Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), showing at least an order of magnitude improvement in energy efficiency versus even our smallest neural network baseline model with a single hidden layer. Our work thus motivates further investigation into molecular representation learning to develop ultra-efficient pre-screening tools. We make our code publicly available at https://github.com/LLNL/hdbind.

59 BASIC BIOLOGICAL SCIENCES↗

Discovery of a metastable Al{sub 20}Sm{sub 4} phase

We present an efficient genetic algorithm, integrated with experimental diffraction data, to solve a nanoscale metastable Al{sub 20}Sm{sub 4} phase that evolves during crystallization of an amorphous magnetron sputtered Al{sub 90}Sm{sub 10} alloy. The excellent match between calculated and experimental X-ray diffraction patterns confirms an accurate description of this metastable phase. Molecular dynamic simulations of crystal growth from the liquid phase predict the formation of disordered defects in the devitrified crystal.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗