Search NASA⌕ Search

SEARCH · Search NASA

Results for “high throughput experiments”

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 145 records · Page 8
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Pose Classification Using Three-Dimensional Atomic Structure-Based Neural Networks Applied to Ion Channel–Ligand Docking

The identification of promising lead compounds showing pharmacological activities toward a biological target is essential in early stage drug discovery. With the recent increase in available small-molecule databases, virtual high-throughput screening using physics-based molecular docking has emerged as an essential tool in assisting fast and cost-efficient lead discovery and optimization. However, the best scored docking poses are often suboptimal, resulting in incorrect screening and chemical property calculation. We address the pose classification problem by leveraging data-driven machine learning approaches to identify correct docking poses from AutoDock Vina and Glide screens. To enable effective classification of docking poses, we present two convolutional neural network approaches: a three-dimensional convolutional neural network (3D-CNN) and an attention-based point cloud network (PCN) trained on the PDBbind refined set. We demonstrate the effectiveness of our proposed classifiers on multiple evaluation data sets including the standard PDBbind CASF-2016 benchmark data set and various compound libraries with structurally different protein targets including an ion channel data set extracted from Protein Data Bank (PDB) and an in-house KCa3.1 inhibitor data set. Our experiments show that excluding false positive docking poses using the proposed classifiers improves virtual high-throughput screening to identify novel molecules against each target protein compared to the initial screen based on the docking scores.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

PlasmidMaker is a versatile, automated, and high throughput end-to-end platform for plasmid construction

Abstract Plasmids are used extensively in basic and applied biology. However, design and construction of plasmids, specifically the ones carrying complex genetic information, remains one of the most time-consuming, labor-intensive, and rate-limiting steps in performing sophisticated biological experiments. Here, we report the development of a versatile, robust, automated end-to-end platform named PlasmidMaker that allows error-free construction of plasmids with virtually any sequences in a high throughput manner. This platform consists of a most versatile DNA assembly method using Pyrococcus furiosus Argonaute ( Pf Ago)-based artificial restriction enzymes, a user-friendly frontend for plasmid design, and a backend that streamlines the workflow and integration with a robotic system. As a proof of concept, we used this platform to generate 101 plasmids from six different species ranging from 5 to 18 kb in size from up to 11 DNA fragments. PlasmidMaker should greatly expand the potential of synthetic biology.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

PlasmidMaker: a Versatile, Automated, and High Throughput End-to-End Platform for Plasmid Construction

Plasmids are used extensively in basic and applied biology. However, design and construction of plasmids, specifically the ones carrying complex genetic information, remains one of the most time-consuming, labor-intensive, and rate-limiting steps in performing sophisticated biological experiments. Here, we report the development of a versatile, robust, automated end-to-end platform named PlasmidMaker that allows error-free construction of plasmids with virtually any sequences in a high-throughput manner. This platform consists of a most versatile DNA assembly method using Pyrococcus furiosus Argonaute ( Pf Ago)-based artificial restriction enzymes, a user-friendly frontend for plasmid design, and a backend that streamlines the workflow and integration with a robotic system. As a proof of concept, we used this platform to generate ~100 plasmids from six different species ranging from 5 to 18 kb in size from up to 11 DNA fragments within 3 days. PlasmidMaker should greatly expand the potential of synthetic biology.

Enghiad, Behnam↗
DOE OSTI2021

A high solids field-to-fuel research pipeline to identify interactions between feedstocks and biofuel production

Abstract Background Environmental factors, such as weather extremes, have the potential to cause adverse effects on plant biomass quality and quantity. Beyond adversely affecting feedstock yield and composition, which have been extensively studied, environmental factors can have detrimental effects on saccharification and fermentation processes in biofuel production. Only a few studies have evaluated the effect of these factors on biomass deconstruction into biofuel and resulting fuel yields. This field-to-fuel evaluation of various feedstocks requires rigorous coordination of pretreatment, enzymatic hydrolysis, and fermentation experiments. A large number of biomass samples, often in limited quantity, are needed to thoroughly understand the effect of environmental conditions on biofuel production. This requires greater processing and analytical throughput of industrially relevant, high solids loading hydrolysates for fermentation, and led to the need for a laboratory-scale high solids experimentation platform. Results A field-to-fuel platform was developed to provide sufficient volumes of high solids loading enzymatic hydrolysate for fermentation. AFEX pretreatment was conducted in custom pretreatment reactors, followed by high solids enzymatic hydrolysis. To accommodate enzymatic hydrolysis of multiple samples, roller bottles were used to overcome the bottlenecks of mixing and reduced sugar yields at high solids loading, while allowing greater sample throughput than possible in bioreactors. The roller bottle method provided 42–47% greater liquefaction compared to the batch shake flask method for the same solids loading. In fermentation experiments, hydrolysates from roller bottles were fermented more rapidly, with greater xylose consumption, but lower final ethanol yields and CO 2 production than hydrolysates generated with shake flasks. The entire platform was tested and was able to replicate patterns of fermentation inhibition previously observed for experiments conducted in larger-scale reactors and bioreactors, showing divergent fermentation patterns for drought and normal year switchgrass hydrolysates. Conclusion A pipeline of small-scale AFEX pretreatment and roller bottle enzymatic hydrolysis was able to provide adequate quantities of hydrolysate for respirometer fermentation experiments and was able to overcome hydrolysis bottlenecks at high solids loading by obtaining greater liquefaction compared to batch shake flask hydrolysis. Thus, the roller bottle method can be effectively utilized to compare divergent feedstocks and diverse process conditions.

09 BIOMASS FUELS↗
DOE OSTI2024

Bragg glass signatures in Pd x ErTe 3 with X-ray diffraction temperature clustering

The Bragg glass phase is a nearly perfect crystal with glassy features predicted to occur in vortex lattices and charge-density-wave systems in the presence of disorder. Detecting it has been challenging, despite its sharp theoretical definition in terms of diverging correlation lengths. Here we present bulk probe evidence supporting a Bragg glass phase in the systematically disordered charge-density-wave material of Pd x ErTe 3 . We do this by using comprehensive X-ray data and a machine-learning-based analysis tool called X-ray diffraction temperature clustering (X-TEC). We establish a diverging correlation length in samples with moderate intercalation over a wide temperature range. To enable this analysis, we introduced a high-throughput measure of inverse correlation length that we call peak spread. The detection of Bragg glass order and the resulting phase diagram advance our understanding of the complex interplay between disorder and fluctuations. Moreover, the use of our analysis technique to target fluctuations through a high-throughput measure of peak spread can revolutionize the study of fluctuations in scattering experiments.

36 MATERIALS SCIENCE↗
DOE OSTI2023

ESnet/JLab FPGA Accelerated Transport

To increase the science rate for high data rates/volumes, Thomas Jefferson National Accelerator Facility (JLab) has partnered with Energy Sciences Network (ESnet) to define an edge to data center traffic shaping / steering transport capability featuring data event aware network shaping and forwarding. The keystone of this ESnet+JLab FPGA Accelerated Transport (EJFAT) is the joint development of an AI/ML directed dynamic compute work Load Balancer (LB) of UDP streamed data. The LB is a suite consisting of a Field Programmable Gate Array (FPGA) executing the dynamically configurable, low fixed latency LB data plane featuring real-time packet redirection and high throughput, and a control plane running on the FPGA host computer that monitors network and compute farm telemetry in order to make dynamic AI/ML guided decisions for destination compute host redirection/load balancing and destination resource provisioning. The LB provides for three-tier horizontal scaling across LB suites, core compute hosts, and CPUs within a host. The LB effectively provides seamless integration of edge/core computing to support direct experimental data processing for immediate use by JLab science programs and others such as the EIC as well as data centers of the future requiring high throughput and low latency for both hot and cooled data for both running experiment data acquisition systems and data center use cases.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2011

NASA Tech Briefs, November 2011

The topics include: 1) Flight Test Results from the Rake Airflow Gage Experiment on the F-15B; 2) Telemetry and Science Data Software System; 3) CropEx Web-Based Agricultural Monitoring and Decision Support; 4) High-Performance Data Analysis Tools for Sun-Earth Connection Missions; 5) Experiment in Onboard Synthetic Aperture Radar Data Processing; 6) Microfabrication of a High-Throughput Nanochannel Delivery/Filtration System; 7) Improved Design and Fabrication of Hydrated-Salt Pills; 8) Monolithic Flexure Pre-Stressed Ultrasonic Horns; 9) Cryogenic Quenching Process for Electronic Part Screening; 10) Broadband Via-Less Microwave Crossover Using Microstrip-CPW Transitions; 11) Wheel-Based Ice Sensors for Road Vehicles; 12) G-DYN Multibody Dynamics Engine; 13) Multibody Simulation Software Testbed for Small-Body Exploration and Sampling; 14) Propulsive Reaction Control System Model; 15) Licklider Transmission Protocol Implementation; 16) Core Recursive Hierarchical Image Segmentation; 17) Two-Stage Centrifugal Fan; 18) Combined Structural and Trajectory Control of Variable-Geometry Planetary Entry Systems; 19) Pressure Regulator With Internal Ejector Circulation Pump, Flow and Pressure Measurement Porting, and Fuel Cell System Integration Options; 20) Temperature-Sensitive Coating Sensor Based on Hematite; 21) Standardization of a Volumetric Displacement Measurement for Two-Body Abrasion Scratch Test Data Analysis; 22) Detection of Carbon Monoxide Using Polymer-Carbon Composite Films; 23) Substituted Quaternary Ammonium Salts Improve Low-Temperature Performance of Double-Layer Capacitors; 24) Sustainably Sourced, Thermally Resistant, Radiation Hard Biopolymer; 25) Integrated Lens Antennas for Multi-Pixel Receivers; 26) 180-GHz Interferometric Imager; 27) Maturation of Structural Health Management Systems for Solid Rocket Motors; 28) Validating Phasing and Geometry of Large Focal Plane Arrays; 29) Transverse Pupil Shifts for Adaptive Optics Non-Common Path Calibration; 30) Qualification of Fiber Optic Cables for Martian Extreme Temperature Environments; 31) Solid-State Spectral Light Source System; 32) Multiple-Event, Single-Photon Counting Imaging Sensor; 33) Surface Modeling to Support Small-Body Spacecraft Exploration and Proximity Operations; and 34) Achieving Exact and Constant Turnaround Ratio in a DDS-Based Coherent Transponder.

Source record↗
DOE OSTI2024

Emerging Jets Search, Triton Server Deployment, and Track Quality Development: Machine Learning Applications in High Energy Physics

Machine learning is becoming prevalent in high energy physics, with numerous applications in physics analyses and event reconstruction showing great improvements compared to traditional computing methods. This thesis studies three projects which each propose new avenues for machine learning applications within the high energy physics CMS experiment located at CERN. In the first project, a search for a dark matter signal called “emerging jets” is performed, using graph neural networks to greatly increase sensitivity to the signal’s signature within the data. The result of this dark matter search sets the most stringent exclusion limits to date on theoretical emerging jet models. Motivated by inefficiencies encountered when processing the emerging jet graph neural network at Fermi National Accelerator Laboratory’s computing centers, the second project re-optimizes the computing centers for machine learning inference. This re-optimization uses NVIDIA Triton Inference Servers to process users’ analysis code heterogeneously, therefore achieving high processing throughput and decreasing user time-to-insight. The last project focuses on an upgrade to the CMS experiment’s real-time event selection system which improves physics object reconstruction under harsh processing conditions. A boosted decision tree is used to quickly and efficiently quantify a reconstructed particle’s “track quality” in order to remove particle tracks reconstructed erroneously. In summary, this thesis will not only present examples of how high energy physics can greatly benefit by leveraging machine learning techniques for physics analysis and reconstruction, but will also provide guidance on how the field can prepare for the inevitable increase in machine learning applications.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Autonomous organic synthesis for redox flow batteries via flexible batch Bayesian optimization

Traditional trial-and-error methods for materials discovery are inefficient to meet the urgent demands posed by the rapid progression of climate change. This urgency has driven the increasing interest in integrating robotics and machine learning into materials research to accelerate experimental learning. However, idealized decision-making frameworks to achieve maximum sampling efficiency are not always compatible with high-throughput experimental workflows inside a laboratory. For multi-step chemical processes, differences in hardware capacities can complicate the digital framework by introducing constraints on the maximum number of samples in each step of the experiment, hence causing varying batch sizes in variable selection within the same batch. Therefore, designing flexible sampling algorithms is necessary to accommodate the multi-step synthesis with practical constraints unique to each high-throughput workflow. In this work, we designed and employed three strategies on a high-throughput robotic platform to optimize the sulfonation reaction of redox-active molecules used in flow batteries. Our strategies adapt to the multi-step experimental workflow, where their formulation and heating steps are separate, causing varying batch size requirements. By strategically sampling using clustering and mixed-variable batch Bayesian optimization, we were able to iteratively identify optimal conditions that maximize the yields. Our work presents a flexible approach that allows tailoring the machine learning decision-making to suit the practical constraints in individual high-throughput experimental platforms, followed by performing resource-efficient yield optimization using available open-source Python libraries.

Tamura, Clara [Univ. of Washington, Seattle, WA (U↗
DOE OSTI2023

Artificial intelligence for materials research at extremes

Abstract Materials development is slow and expensive, taking decades from inception to fielding. For materials research at extremes, the situation is even more demanding, as the desired property combinations such as strength and oxidation resistance can have complex interactions. Here, we explore the role of AI and autonomous experimentation (AE) in the process of understanding and developing materials for extreme and coupled environments. AI is important in understanding materials under extremes due to the highly demanding and unique cases these environments represent. Materials are pushed to their limits in ways that, for example, equilibrium phase diagrams cannot describe. Often, multiple physical phenomena compete to determine the material response. Further, validation is often difficult or impossible. AI can help bridge these gaps, providing heuristic but valuable links between materials properties and performance under extreme conditions. We explore the potential advantages of AE along with decision strategies. In particular, we consider the problem of deciding between low-fidelity, inexpensive experiments and high-fidelity, expensive experiments. The cost of experiments is described in terms of the speed and throughput of automated experiments, contrasted with the human resources needed to execute manual experiments. We also consider the cost and benefits of modeling and simulation to further materials understanding, along with characterization of materials under extreme environments in the AE loop. Graphical abstract AI sequential decision-making methods for materials research: Active learning, which focuses on exploration by sampling uncertain regions, Bayesian and bandit optimization as well as reinforcement learning (RL), which trades off exploration of uncertain regions with exploitation of optimum function value. Bayesian and bandit optimization focus on finding the optimal value of the function at each step or cumulatively over the entire steps, respectively, whereas RL considers cumulative value of the labeling function, where the latter can change depending on the state of the system (blue, orange, or green).

36 MATERIALS SCIENCE↗
NASA NTRSDate not supplied

NASA Open Science Data Repository: Maximizing Spaceflight Bioscience Data

The next era in human space exploration is rapidly approaching and will require the use of countermeasures to deep space health hazards. The development of countermeasures (or, the re-purposing of existing agents) will be highly dependent on our understanding of basic biological responses to space stressors (e.g. ionizing radiation, altered gravitational fields, altered day-night cycles, confinement, isolation, hostile-closed environments, distance-duration from Earth, exposure to celestial regolith, etc.). The fast-growing array of space biological data, which in the past was simply archived after minimal analysis, holds great potential if it can be reorganized and formatted for data re-analysis and re-use via Open Science. Organizing the data for such analysis is a challenge because of its diverse nature (molecular, cellular, tissue, imaging, whole organism and behavior). To address the challenges posed by gaining new knowledge from a vast and diverse amount of biological, health and environmental data in space, the NASA Open Science Data Repository (OSDR - osdr.nasa.gov/bio) plays a crucial role in curating and openly publishing biological data from space-related experiments. Its design incorporates successes and lessons from NASA GeneLab, encompassing not only high-throughput sequencing data but also physiological, phenotypic, and telemetry data. The OSDR makes space biological data FAIR (findable, accessible, interoperable, reusable), and facilitates effective data ingestion, dissemination, and Open Science collaborations. The OSDR also has the capability to integrate human astronaut data with state-of-the-art security and accessibility procedures. We will discuss here several strategies that NASA’s Biological and Physical Science Division have put in place to maximize the return on investment for spaceflight bioscience data.

space biology↗
DOE OSTI2024

Adaptable SEC‐SAXS data collection for higher quality structure analysis in solution

Abstract The two major challenges in synchrotron size‐exclusion chromatography coupled in‐line with small‐angle x‐ray scattering (SEC‐SAXS) experiments are the overlapping peaks in the elution profile and the fouling of radiation‐damaged materials on the walls of the sample cell. In recent years, many post‐experimental analyses techniques have been developed and applied to extract scattering profiles from these problematic SEC‐SAXS data. Here, we present three modes of data collection at the BioSAXS Beamline 4–2 of the Stanford Synchrotron Radiation Lightsource (SSRL BL4‐2). The first mode, the High‐Resolution mode, enables SEC‐SAXS data collection with excellent sample separation and virtually no additional peak broadening from the UHPLC UV detector to the x‐ray position by taking advantage of the low system dispersion of the UHPLC. The small bed volume of the analytical SEC column minimizes sample dilution in the column and facilitates data collection at higher sample concentrations with excellent sample economy equal to or even less than that of the conventional equilibrium SAXS method. Radiation damage problems during SEC‐SAXS data collection are evaded by additional cleaning of the sample cell after buffer data collection and avoidance of unnecessary exposures through the use of the x‐ray shutter control options, allowing sample data collection with a clean sample cell. Therefore, accurate background subtraction can be performed at a level equivalent to the conventional equilibrium SAXS method without requiring baseline correction, thereby leading to more reliable downstream structural analysis and quicker access to new science. The two other data collection modes, the High‐Throughput mode and the Co‐Flow mode, add agility to the planning and execution of experiments to efficiently achieve the user's scientific objectives at the SSRL BL4‐2.

Matsui, Tsutomu↗
DOE OSTI2023

Fast 2D Bicephalous Convolutional Autoencoder for Compressing 3D Time Projection Chamber Data

High-energy large-scale particle colliders produce data at high speed in the order of 1 terabytes per second in nuclear physics and petabytes per second in high energy physics. Developing real-time data compression algorithms to reduce such data at high throughput to fit permanent storage has drawn increasing attention. Specifically, at the newly constructed sPHENIX experiment at the Relativistic Heavy Ion Collider (RHIC), a time projection chamber is used as the main tracking detector, which records particle trajectories in a volume of three-dimensional (3D) cylinder. The resulting data are usually very sparse with occupancy around 10.8%. Such sparsity presents a challenge to conventional learning-free lossy compression algorithms, such as SZ, ZFP, and MGARD. The 3D convolutional neural network (CNN)-based approach, Bicephalous Convolutional Autoencoder (BCAE), outperforms traditional methods both in compression rate and reconstruction accuracy. BCAE can also utilize the computation power of graphical processing units suitable for deployment in a modern heterogeneous highperformance computing environment. This work introduces two BCAE variants: BCAE++ and BCAE-2D. BCAE++ achieves a 15% better compression ratio and a 77% better reconstruction accuracy measured in mean absolute error compared with BCAE. BCAE-2D treats the radial direction as the channel dimension of an image, resulting in a 3× speedup in compression throughput. In addition, we demonstrate an unbalanced autoencoder with a larger decoder can improve reconstruction accuracy without significantly sacrificing throughput. Lastly, we observe both the BCAE++ and BCAE-2D can benefit more from using half-precision mode in throughput (76 - 79% increase) without loss in reconstruction accuracy. The source code and links to data and pretrained models can be found at https://github.com/BNL-DAQ-LDRD/NeuralCompression_v2

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2022

HT-SIP: a semi-automated stable isotope probing pipeline identifies cross-kingdom interactions in the hyphosphere of arbuscular mycorrhizal fungi

Abstract Background Linking the identity of wild microbes with their ecophysiological traits and environmental functions is a key ambition for microbial ecologists. Of many techniques that strive for this goal, Stable-isotope probing—SIP—remains among the most comprehensive for studying whole microbial communities in situ. In DNA-SIP, actively growing microorganisms that take up an isotopically heavy substrate build heavier DNA, which can be partitioned by density into multiple fractions and sequenced. However, SIP is relatively low throughput and requires significant hands-on labor. We designed and tested a semi-automated, high-throughput SIP (HT-SIP) pipeline to support well-replicated, temporally resolved amplicon and metagenomics experiments. We applied this pipeline to a soil microhabitat with significant ecological importance—the hyphosphere zone surrounding arbuscular mycorrhizal fungal (AMF) hyphae. AMF form symbiotic relationships with most plant species and play key roles in terrestrial nutrient and carbon cycling. Results Our HT-SIP pipeline for fractionation, cleanup, and nucleic acid quantification of density gradients requires one-sixth of the hands-on labor compared to manual SIP and allows 16 samples to be processed simultaneously. Automated density fractionation increased the reproducibility of SIP gradients compared to manual fractionation, and we show adding a non-ionic detergent to the gradient buffer improved SIP DNA recovery. We applied HT-SIP to 13 C-AMF hyphosphere DNA from a 13 CO 2 plant labeling study and created metagenome-assembled genomes (MAGs) using high-resolution SIP metagenomics (14 metagenomes per gradient). SIP confirmed the AMF Rhizophagus intraradices and associated MAGs were highly enriched (10–33 atom% 13 C), even though the soils’ overall enrichment was low (1.8 atom% 13 C). We assembled 212 13 C-hyphosphere MAGs; the hyphosphere taxa that assimilated the most AMF-derived 13 C were from the phyla Myxococcota, Fibrobacterota, Verrucomicrobiota, and the ammonia-oxidizing archaeon genus Nitrososphaera . Conclusions Our semi-automated HT-SIP approach decreases operator time and improves reproducibility by targeting the most labor-intensive steps of SIP—fraction collection and cleanup. We illustrate this approach in a unique and understudied soil microhabitat—generating MAGs of actively growing microbes living in the AMF hyphosphere (without plant roots). The MAGs’ phylogenetic composition and gene content suggest predation, decomposition, and ammonia oxidation may be key processes in hyphosphere nutrient cycling.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

High-Temperature SiC Cladding End Plug Irradiation Design and HFIR Readiness

This report documents the design of a High Flux Isotope Reactor (HFIR) irradiation experiment intended to evaluate irradiation effects on the hermeticity of silicon carbide (SiC) end plug specimens under a radial fast neutron flux gradient at representative light-water reactor (LWR) temperatures of approximately 300 °C. The overarching goal of this work is to statistically evaluate SiC end plug hermeticity and mechanical properties following irradiation using the high-throughput irradiation capability discussed here. Each specimen consists of a short section of SiC fiber–reinforced SiC (SiC/SiC) tube with a single monolithic SiC end plug joined to one end. The experiment allows up to 66 specimens to be irradiated in six different stacks within a dry, sealed irradiation capsule derived from the previously developed high-temperature SiC/SiC cladding bowing experiment. The neutronics basis, thermal analysis, and HFIR readiness of the experiment are discussed in this report for two possible design cases. The first design case is based on existing approval documentation and components that are on hand and approved for use, so the experiment insertion would require only specimen receipt, specimen pre-irradiation characterization, experiment assembly, and final fabrication package approval. The second design case provides improved thermal robustness and the preferred end plug geometry but requires fabrication of a modified holder and revisions to the HFIR approval documentation, in addition to the other activities required for the first design case, before insertion.

Hott, Daniel [Oak Ridge National Laboratory (ORNL)↗
DOE OSTI2023

Hybrid Data‐Driven Discovery of High‐Performance Silver Selenide‐Based Thermoelectric Composites

Optimizing material compositions often enhances thermoelectric performances. However, the large selection of possible base elements and dopants results in a vast composition design space that is too large to systematically search using solely domain knowledge. To address this challenge, a hybrid data-driven strategy that integrates Bayesian optimization (BO) and Gaussian process regression (GPR) is proposed to optimize the composition of five elements (Ag, Se, S, Cu, and Te) in AgSe-based thermoelectric materials. Data is collected from the literature to provide prior knowledge for the initial GPR model, which is updated by actively collected experimental data during the iteration between BO and experiments. Within seven iterations, the optimized AgSe-based materials prepared using a simple high-throughput ink mixing and blade coating method deliver a high power factor of 2100 µW m −1 K −2 , which is a 75% improvement from the baseline composite (nominal composition of Ag 2 Se 1 ). In conclusion, the success of this study provides opportunities to generalize the demonstrated active machine learning technique to accelerate the development and optimization of a wide range of material systems with reduced experimental trials.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Use of Graph Theory and Neural Networks for Microstructural Classification

Recent advances in materials data analytics have provided new avenues for determining process-structure-property (PSP) linkages in a variety of materials. Machine learning techniques including few-shot learning have increased the efficiency of classifying microscopy images for the purposes of material characterization. Modifications in segmentation also show potential in improving the accuracy of our current pyCHIP classifier. Replacing previous encoders trained on ImageNet with those trained on microscopy images like MicroNet has initially shown better performance at classifying images of irradiated samples. Additionally, different normalization approaches were tested to show no discernable effect on classification. The Louvain method for community detection is analyzed on a set of irradiated samples with different parameters to determine which proved beneficial under what circumstances. We suggest that microscopy experiments be automated in the future using a combination of these techniques to enable high-throughput analyses.

36 MATERIALS SCIENCE↗