Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data diversity”

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 1,639 records · Page 91

Exascale Virtualized and Programmable Distributed Cyber Resource Control: Final Scientific Technical Report

Extreme-science drives the need for distributed exascale processing and communications that are carefully, yet flexibly, managed. Exponential growth of data for scientific simulations, experimental data, collaborative data analyses, remote visualization and GRID computing requirements of scientists in fields as diverse as high energy physics, climate change, genomics, fusion, synchrotron radiation, material science, medicine, and other scientific disciplines cannot be accommodated by simply applying existing transport protocols to faster pipes. Further, scientific challenges today demand diverse research teams, heightening the need for and increasing the complexity of collaboration. To address these issues within the network layer and physical layer, we have performed a number of research activities surrounding effective allocation and management of elastic optical network (EON) resources, particularly focusing on FlexGrid transponders. FlexGrid transponders support the opportunity to build Layer-1 connections at a wide range of bandwidths and to reconfigure them rapidly. The new flexibility supports complex new ways of using the physical layer that must be carefully managed and hidden from the scientist end-users. FlexGrid networks utilize flexible (or elastic) spectral bandwidths for each data link without using fixed wavelength grids. The flexibility in spectrum allocation brings many appealing features to network operations. Current networks are designed for the worst case impairments in transmission performance and the assigned spectrum is over-provisioned. In contrast, the FlexGrid networks can operate with the highest spectral efficiency and minimum bandwidth for the given traffic demand while meeting the minimum quality of transmission (QoT) requirement. Two primary focuses of our research are: (1) resource and spectrum allocation (RSA) for IP traffic over EONs, and (2) RSA for cross-domain optical networks. Previous work concentrates primarily on large file transfers within a single domain. Adding support for IP traffic changes the nature of the RSA problem: instead of choosing to accept or deny each request for network support, IP traffic is inherently elastic and thus lends itself to a bandwidth maximization formulation. We developed a number of algorithms that could be easily deployed within existing and new FlexGrid networks, leading to networks that better support scientific collaboration. Cross-domain RSA research is essential to support large-scale FlexGrid networks, since configuration information is generally not shared or coordinated across domains. The results presented here are in their early stages. They are technically feasible and practical, but still require coordination among organizations and equipment owners and a higher-layer framework for managing network requests.

42 ENGINEERING↗

Toward Real Time Data Analysis for Smart Grids

This paper describes the architecture and design of a novel system for supporting large-scale real-time data analysis for future power grid systems. The widespread deployment of renewable generation, smart grid controls, energy storage, plug-in hybrids, and new conducting materials will require fundamental changes in the operational concepts and principal components of the grid. As a result, the whole system becomes highly dynamic and requires constant adjusting based on real time data. Even though millions of sensors such as phase measurement units (PMU) and smart meters are being widely deployed, a data layer that can analyze this amount of data in real time is needed. Unlike the data fabric in other cloud services, the data layer for smart grids has some unique design requirements. First, this layer must provide real time guarantees. Second, this layer must be scalable to allow a large number of applications to access the data from millions of sensors in real time. Third, reliability is critical and this layer must be able to continue to provide service in face of failures. Fourth, this layer must be secure. We address these challenges though a scalable system architecture that integrates the I/O and data processing capability in a devise set of devices. Data process operations can be placed anywhere from sensors, data storage devices, to control centers. We further employ compression to improve performance. We design a lightweight compression customized for power grid data. Our system can reduce end-to-end response time by reduce I/O overhead through compression and overlap compression operations with I/O. The initial prototype of our system was demonstrated with several use cases from PNNL’s FPGI and show that our system can provide real time guarantees to a diverse set of applications.

Data Analytics, Power Grid, Big Data↗

Continental Spatio-Temporal Data Analysis with Linear Spectral Mixture Model Using FOSS

This work demonstrates the development and implementation of a Fully Constrained Least Squares (FCLS) unmixing model developed in C++ programming language with OpenCV package and boost C++ libraries in the NASA Earth Exchange (NEX). Visualization of the results is supported by GRASS GIS and statistical analysis is carried in R in a Linux system environment. FCLS was first tested on computer simulated data with Gaussian noise of various signal-to-noise ratio, and Landsat data of an agricultural scenario and an urban environment using a set of global end members of substrate (soils, sediments, rocks, and non-photosynthetic vegetation), vegetation that includes green photosynthetic plants and dark objects which encompasses absorptive substrate materials, clear water, deep shadows, etc. For the agricultural scenario, a spectrally diverse collection of 11 scenes of Level 1 terrain corrected, cloud free Landsat-5 TM data of Fresno, California, USA were unmixed and the results were validated with the corresponding ground data. To study an urbanized landscape, a clear sky Landsat-5 TM data were unmixed and validated with coincident World View-2 abundance maps (of 2 m spatial resolution) for an area of San Francisco, California, USA. The results were evaluated using descriptive statistics, correlation coefficient, RMSE, probability of success, boxplot and bivariate distribution function. Finally, FCLS was used for sub-pixel land cover analysis of the monthly WELD (Wen-enabled Landsat data) repository from 2008 to 2011 of North America. The abundance maps in conjunction with DMSP-OLS nighttime lights data were used to extract the urban land cover features and analyze their spatial-temporal growth.

Landsat Satellites↗

A Conserved Rubredoxin Is Necessary for Photosystem II Accumulation in Diverse Oxygenic Photoautotrophs

In oxygenic photosynthesis, two photosystems work in tandem to harvest light energy and generate NADPH and ATP. Photosystem II(PSII),theprotein-pigment complexthatuseslight energyto catalyze the splitting of water, is assembled from its component parts in a tightly regulated process that requires a number of assembly factors. The 2pac mutant of the unicellular green alga Chlamydomonas reinhardtii was isolated and found to have no detectable PSII activity, whereas other components of the photosynthetic electron transport chain, including photosystem I, were stillfunctional.PSII activitywasfully restoredby complementation with the RBD1 gene, which encodes a small iron-sulfur protein known as a rubredoxin. Phylogenetic evidence supports the hypothesisthatthis rubredoxin anditsorthologs areuniquetooxygenic phototrophs and distinct from rubredoxins in Archaea and bacteria (excluding cyanobacteria). Knockouts of the rubredoxin orthologs in the cyanobacterium Synechocystis sp. PCC 6803 and the plant Arabidopsis thaliana were also found to be specifically affected in PSII accumulation. Taken together, our data suggest that this rubredoxin is necessary for normal PSII activity in a diverse set of organisms that perform oxygenic photosynthesis.

59 BASIC BIOLOGICAL SCIENCES↗

A haplotype‐resolved reference genome of Quercus alba sheds light on the evolutionary history of oaks

Summary White oak ( Quercus alba ) is an abundant forest tree species across eastern North America that is ecologically, culturally, and economically important. We report the first haplotype‐resolved chromosome‐scale genome assembly of Q. alba and conduct comparative analyses of genome structure and gene content against other published Fagaceae genomes. We investigate the genetic diversity of this widespread species and the phylogenetic relationships among oaks using whole genome data. Despite strongly conserved chromosome synteny and genome size across Quercus , certain gene families have undergone rapid changes in size, including defense genes. Unbiased annotation of resistance (R) genes across oaks revealed that the overall number of R genes is similar across species – as are the chromosomal locations of R gene clusters – but, gene number within clusters is more labile. We found that Q. alba has high genetic diversity, much of which predates its divergence from other oaks and likely impacts divergence time estimations. Our phylogenetic results highlight widespread phylogenetic discordance across the genus. The white oak genome represents a major new resource for studying genome diversity and evolution in Quercus . Additionally, we show that unbiased gene annotation is key to accurately assessing R gene evolution in Quercus .

Larson, Drew A. [Department of Biology Indiana Uni↗

Species interactions amplify functional group responses to elevated CO 2 and N enrichment in a 24‐year grassland experiment

Abstract Plant functional groups (FGs) differ in their response to global changes, although species within those groups also vary in such responses. Both species and FG responses to global change are likely influenced by species interactions such as inter‐specific competition and facilitation, which are prevalent in species mixtures but not monocultures. As most studies focus on responses of plants growing in either monocultures or mixtures, but rarely both, it remains unclear how interspecific interactions in diverse ecological communities, especially among species in different FGs, modify FG responses to global changes. To address these issues, we leveraged data from a 16‐species, 24‐year perennial grassland experiment to examine plant FG biomass responses to atmospheric CO 2 , and N inputs at different planted diversity. FGs differed in their responses to N and CO 2 treatments in monocultures. Such differences were amplified in mixtures, where N enrichment strongly increased C3 grass success at ambient CO 2 and C4 grass success at elevated CO 2 . Legumes declined with N enrichment in mixtures at both CO 2 levels and increased with elevated CO 2 in the initial years of the experiment. Our results suggest that previous studies that considered responses to global changes in monocultures may underestimate biomass changes in diverse communities where interspecific interactions can amplify responses. Such effects of interspecific interactions on responses of FGs to global change may impact community composition over time and consequently influence ecosystem functions.

Mohanbabu, Neha↗

The Western Lake Erie culture collection: A promising resource for evaluating the physiological and genetic diversity of Microcystis and its associated microbiome

Cyanobacterial harmful algal blooms (cyanoHABs) dominated by Microcystis spp. have significant public health and economic implications in freshwater bodies around the world. These blooms are capable of producing a variety of cyanotoxins, including microcystins, that affect fishing and tourism industries, human and environmental health, and access to drinking water. In this study, we isolated and sequenced the genomes of 21 primarily unialgal Microcystis cultures collected from western Lake Erie between 2017 and 2019. While some cultures isolated in different years have a high degree of genetic similarity (genomic Average Nucleotide Identity >99%), genomic data show that these cultures also represent much of the breadth of known Microcystis diversity in natural populations. Only five isolates contained all the genes required for microcystin biosynthesis while two isolates contained a previously described partial mcy operon. Microcystin production within cultures was also assessed using Enzyme-Linked Immunosorbent Assay (ELISA) and supported genomic results with high concentrations (up to 900 μg L⁻¹) in cultures with complete mcy operons and no or low toxin detected otherwise. These xenic cultures also contained a substantial diversity of bacteria associated with Microcystis, which has become increasingly recognized as an essential component of cyanoHAB community dynamics. These results highlight the genomic diversity among Microcystis strains and associated bacteria in Lake Erie, and their potential impacts on bloom development, toxin production, and toxin degradation. Furthermore, this culture collection significantly increases the availability of environmentally relevant Microcystis strains from temperate North America.

CyanoHABs↗

HydroForecast Long-term: Improving hydropower’s resilience to climate change through accurate climate-scale

With hydrologic patterns and water availability across the globe shifting due to climate change, advancements in hydrologic prediction systems can help significantly reduce the uncertainties that utilities and water supply entities have in their decision making. Understanding and estimating hydrology at the climate scale is critical for managing water resources under changing climate scenarios. This project focuses on integrating state-of-the-art neural network modeling with downscaled climate projections to deliver the reliable water supply projections decades into the future to meet an urgent need from hydropower operators and water utilities. In this Phase 1 DOE SBIR proposal, we developed and validated a theory-guided neural network model, HydroForecast Long-term, for climate-scale hydrology and implemented the model within existing HydroForecast infrastructure. HydroForecast Long-term combines the most accurate streamflow modeling system with a flexible and scalable data architecture to generate water supply projections out to the year 2100. This report illustrates that we have achieved our four objectives: 1) create a prototype of HydroForecast Long-term, building the neural network prediction model, 2) build an automated data input pipeline that processes large amounts of data from the latest global temperature and precipitation climate models; 3) benchmark the accuracy of the hydrologic model over the recent two decades over a large set of diverse basins, and 4) create a set of output visuals and summary metrics informed by customer feedback that connect the data to critical decision points. This work empowers water users to make data-informed decisions supporting a resilient, renewable-powered grid and water system. The results advance the Department of Energy’s mission by addressing critical gaps in water supply planning under climate change.

13 HYDRO ENERGY↗

Lineage-based functional types: characterising functional diversity to enhance the representation of ecological behaviour in Land Surface Models

Process-based vegetation models attempt to represent the wide range of trait variation in biomes by grouping ecologically similar species into plant functional types (PFTs). This approach has been successful in representing many aspects of plant physiology and biophysics but struggles to capture biogeographic history and ecological dynamics that determine biome boundaries and plant distributions. Grass-dominated ecosystems are broadly distributed across all vegetated continents and harbour large functional diversity, yet most Land Surface Models (LSMs) summarise grasses into two generic PFTs based primarily on differences between temperate C 3 grasses and (sub)tropical C 4 grasses. Incorporation of species-level trait variation is an active area of research to enhance the ecological realism of PFTs, which form the basis for vegetation processes and dynamics in LSMs. Using reported measurements, we developed grass functional trait values (physiological, structural, biochemical, anatomical, phenological, and disturbance-related) of dominant lineages to improve LSM representations. In this work, our method is fundamentally different from previous efforts, as it uses phylogenetic relatedness to create lineage-based functional types (LFTs), situated between species-level trait data and PFT-level abstractions, thus providing a realistic representation of functional diversity and opening the door to the development of new vegetation models.

54 ENVIRONMENTAL SCIENCES↗

Differences in urban plant community compositions across an urban-rural gradient in Knoxville, TN

Urban forests, or vegetation in areas under heavy human influence, provide many ecosystem services to urban residents such as localized cooling via evapotranspiration, shade, filtering of air pollution, and the associated health benefits of natural spaces. In order to quantify the magnitude of localized cooling by trees growing in varying levels of urbanization (based on % impervious surfaces, e.g., buildings, pavement), urban forest species composition, tree size, and tree density must be characterized. As a part of Oak Ridge National Laboratory’s (ORNL) urban forest temperature study, we conducted tree censuses in five Knoxville city parks where ORNL meteorological stations are deployed. Moreover, we measured every woody plant ≥ 5 cm diameter at breast height (DBH) within a 50 m radius of each site’s meteorological station for its DBH and species identification. When possible, individuals were identified down to species. Certain genera (Quercus spp., Carya spp., Pinus spp.) were identified down to genera in interest of time. Individual and total site basal area were calculated from measured DBH data. Results show notable differences in urban plant community compositions and total woody plant basal area across sites, with more urban sites closer to downtown (West View and SEEED) having lower tree basal area than the more suburban sites (West Hills, Cumberland Estates, and Victor Ashe). We identified 54 species across all sites, with West Hills and Victor Ashe having the highest species diversity. Our results show differences in forest compositions and sizes across Knoxville, which are currently informing ORNL’s evapotranspiration estimates for each site. Data Summary: Census data for West Hills (WH), Cumberland Estates (CE), Victor Ashe (VA), West View (WV), and Socially Equal Energy Efficient Development or SEEED (SD) urban forests in Knoxville, TN, USA, including tree size based on diameter at breast height (DBH; 1.3 m), species identification (Latin and common names), and basal area per stem (BA=π×[.5*DBH]^2). Field data are summarized in this file: “Community_Composition_Data.CSV”. Site-specific data detailing each site’s coordinates, number of stems measured at DBH, average tree DBH, α-diversity (number of species present), and total site basal area (sum of individual basal areas per site) are in this file: “Site_Comparisons.CSV”.

Warren, Jeffrey [ORNL] (ORCID:0000000206804697)↗

A genomic view of Earth’s biomes

Microorganisms are essential to all life on Earth through critical roles in key biological processes and diverse interactions with other organisms that shape ecosystems, drive biogeochemical cycles and influence both human health and environmental health. High-throughput sequencing from environmental samples has revolutionized the understanding of microbial diversity and functions. With vast amounts of genomes now available across Earth’s biomes, these data provide a blueprint of microbial life that can be harnessed for a more holistic understanding of microbiome structure and function across the various ecosystems on Earth. Here we review the application of genome-centric approaches, including recent advances in single-cell sequencing and functional profiling, to survey microbial and viral diversity. Furthermore, we highlight some of the most impactful evolutionary and functional discoveries, explore the spatial diversity and temporal dynamics of microorganisms across diverse environments, and discuss genome-enabled insights into host-associated microorganisms.

Ecology↗

Quantification of the epitope diversity of HIV-1-specific binding antibodies by peptide microarrays for global HIV-1 vaccine development

An effective vaccine against human immunodeficiency virus type 1 (HIV-1) will have to provide protection against a vast array of different HIV-1 strains. Current methods to measure HIV-1-specific binding antibodies following immunization typically focus on determining the magnitude of antibody responses, but the epitope diversity of antibody responses has remained largely unexplored. Here we describe the development of a global HIV-1 peptide microarray that contains 6564 peptides from across the HIV-1 proteome and covers the majority of HIV-1 sequences in the Los Alamos National Laboratory global HIV-1 sequence database. Using this microarray, we quantified the magnitude, breadth, and depth of IgG binding to linear HIV-1 sequences in HIV-1-infected humans and HIV-1-vaccinated humans, rhesus monkeys and guinea pigs. The microarray measured potentially important differences in antibody epitope diversity, particularly regarding the depth of epitope variants recognized at each binding site. Our data suggest that the global HIV-1 peptide microarray may be a useful tool for both preclinical and clinical HIV-1 research.

60 APPLIED LIFE SCIENCES↗

Machine Learning Predictions of Simulated Self-Diffusion Coefficients for Bulk and Confined Pure Liquids

Diffusion properties of bulk fluids have been predicted using empirical expressions and machine learning (ML) models, suggesting that predictions of diffusion also should be possible for fluids in confined environments. The ability to quickly and accurately predict diffusion in porous materials would enable new discoveries and spur development in relevant technologies such as separations, catalysis, batteries, and subsurface applications. Here in this work, we apply artificial neural network (ANN) models to predict the simulated self-diffusion coefficients of real liquids in both bulk and pore environments. The training data sets were generated from molecular dynamics (MD) simulations of Lennard-Jones particles representing a diverse set of 14 molecules ranging from ammonia to dodecane over a range of liquid pressures and temperatures. Planar, cylindrical, and hexagonal pore models consisted of walls composed of carbon atoms. Our simple model for these liquids was primarily used to generate ANN training data, but the simulated self-diffusion coefficients of bulk liquids show excellent agreement with experimental diffusion coefficients. ANN models based on simple descriptors accurately reproduced the MD diffusion data for both bulk and confined liquids, including the trend of increased mobility in large pores relative to the corresponding bulk liquid.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

iVirus: facilitating new insights in viral ecology with software and community data sets imbedded in a cyberinfrastructure

Microbes affect nutrient and energy transformations throughout the world’s ecosystems, yet they do so under viral constraints. In complex communities, viral metagenome (virome) sequencing is transforming our ability to quantify viral diversity and impacts. Although some bottlenecks, for example, few reference genomes and nonquantitative viromics, have been overcome, the void of centralized data sets and specialized tools now prevents viromics from being broadly applied to answer fundamental ecological questions. Here we present iVirus, a community resource that leverages the CyVerse cyberinfrastructure to provide access to viromic tools and data sets. The iVirus Data Commons contains both raw and processed data from 1866 samples and 73 projects derived from global ocean expeditions, as well as existing and legacy public repositories. Through the CyVerse Discovery Environment, users can interrogate these data sets using existing analytical tools (software applications known as ‘Apps’) for assembly, open reading frame prediction and annotation, as well as several new Apps specifically developed for analyzing viromes. Because Apps are web based and powered by CyVerse supercomputing resources, they enable scalable analyses for a broad user base. Finally, a use-case scenario documents how to apply these advances toward new data. This growing iVirus resource should help researchers utilize viromics as yet another tool to elucidate viral roles in nature.

59 BASIC BIOLOGICAL SCIENCES↗

Integrating Radar, Multispectral, and Landing Site Data for Analysis of the Lunar Surface

Radar maps of the Moon have been produced since the late 1960s, and have been used by a number of authors to study the surface roughness, subsurface rock abundance, and dielectric properties of the lunar surface. These studies focused on a range of topics, including the depth and rock population of the regolith, crater ejecta blankets, pyroclastic mantling layers, and cryptomare deposits. Limited radar sounding data from the Apollo missions identified layering in some regions of the maria. As radar datasets have improved in resolution and calibration, it has become more possible to make quantitative comparisons between the backscatter properties of the Moon, other remote-sensing observations, and the ground truth provided by Surveyor photos and Apollo traverses. This presentation will focus on the results of recent studies of the lunar regolith that make use of these diverse sources of information, and discuss research directions that will be possible with radar data to be collected in the near future.

Campbell, B. A.↗

Refining Planetary Boundary Layer Height Retrievals From Micropulse‐Lidar at Multiple ARM Sites Around the World

Abstract Knowledge of the planetary boundary layer height (PBLH) is crucial for various applications in atmospheric and environmental sciences. Lidar measurements are frequently used to monitor the evolution of the PBLH, providing more frequent observations than traditional radiosonde‐based methods. However, lidar‐derived PBLH estimates have substantial uncertainties, contingent upon the retrieval algorithm used. In addressing this, we applied the Different Thermo‐Dynamic Stabilities (DTDS) algorithm to establish a PBLH data set at five separate Department of Energy's Atmospheric Radiation Measurement sites across the globe. Both the PBLH methodology and the products are subject to rigorous assessments in terms of their uncertainties and constraints, juxtaposing them with other products. The DTDS‐derived product consistently aligns with radiosonde PBLH estimates, with correlation coefficients exceeding 0.77 across all sites. This study delves into a detailed examination of the strengths and limitations of PBLH data sets with respect to both radiosonde‐derived and other lidar‐based estimates of the PBLH by exploring their respective errors and uncertainties. It is found that varying techniques and definitions can lead to diverse PBLH retrievals due to the inherent intricacy and variability of the boundary layer. Our DTDS‐derived PBLH data set outperforms existing products derived from ceilometer data, offering a more precise representation of the PBLH. This extensive data set paves the way for advanced studies and an improved understanding of boundary‐layer dynamics, with valuable applications in weather forecasting, climate modeling, and environmental studies.

54 ENVIRONMENTAL SCIENCES↗

Antarctic lake viromes reveal potential virus associated influences on nutrient cycling in ice-covered lakes

The McMurdo Dry Valleys (MDVs) of Antarctica are a mosaic of extreme habitats which are dominated by microbial life. The MDVs include glacial melt holes, streams, lakes, and soils, which are interconnected through the transfer of energy and flux of inorganic and organic material via wind and hydrology. For the first time, we provide new data on the viral community structure and function in the MDVs through metagenomics of the planktonic and benthic mat communities of Lakes Bonney and Fryxell. Viral taxonomic diversity was compared across lakes and ecological function was investigated by characterizing auxiliary metabolic genes (AMGs) and predicting viral hosts. Our data suggest that viral communities differed between the lakes and among sites: these differences were connected to microbial host communities. AMGs were associated with the potential augmentation of multiple biogeochemical processes in host, most notably with phosphorus acquisition, organic nitrogen acquisition, sulfur oxidation, and photosynthesis. Viral genome abundances containing AMGs differed between the lakes and microbial mats, indicating site specialization. Using procrustes analysis, we also identified significant coupling between viral and bacterial communities (p = 0.001). Finally, host predictions indicate viral host preference among the assembled viromes. Collectively, our data show that: (i) viruses are uniquely distributed through the McMurdo Dry Valley lakes, (ii) their AMGs can contribute to overcoming host nutrient limitation and, (iii) viral and bacterial MDV communities are tightly coupled.

Microbiology↗

A Data-Agnostic, Continuous Machine Learning Framework for Application in High Energy Physics and Beyond: Phase 1 Final Scientific/Technical Report

This Phase 1 effort has focused on the development of continual learning frameworks for use in machine learning, specifically in the applied context of High Energy Physics (HEP). Machine learning (ML) is a transformative technology by which computers, typically through the use of neural networks, are able to perform tasks with proficiency that rivals or surpasses that of human users. Model Degradation & Catastrophic Forgetting are two undesired phenomena which can occur in ML where the performance of a model degrades when either deployed on novel data streams, or trained on novel data which are sufficiently different than the data the models were initially trained on. A natural example where these sorts of effects can be observed is in the performance of detectors in harsh environments, where the detector signature may change over the lifetime of the detector as it ages and deteriorates — precisely what occurs in the experiments conducted in HEP. Real world HEP data is therefore an excellent test-ground and use-case for Continual Learning paradigms, which are techniques used in ML to counteract these problems. Ensemble learning is one such technique, where multiple smaller models are trained on subsets of the overall data and are ensembled together during inference. The intuition behind this technique is that, although there are shifts in the distributions which govern the incoming data streams, these shifts are not expected to be homogeneous or global. If a sufficient diversity in solutions within the various sub-models has been achieved, then at least one sub-model is expected to retain its performance within the overall ensemble. One further strength of this approach is that the architectures of the various models do not need to be identical, and in fact even different modalities of data can naturally be combined in this way. This work focused on applying ensemble learning techniques to derive results using two main datasets, anomaly detection in HEP data & time-series forecasting in semiconductor manufacturing data. Semiconductor manufacturing involves data with surprising similarity to that of HEP (e.g. wafer maps look very similar to digi-occupancy maps) and Cerium Lab’s prominence within the semiconductor industry makes semiconductor manufacturing a natural opportunity for commercialization of this work. Our efforts have led to two strong results. The first is that we evaluated the proposed ensembling techniques using previously proposed machine learning architectures for use in anomaly detection, namely AutoEncoder based models and their derivatives. We also developed new architectures which have not been evaluated in this context before. In fact, this work marks the first use of Vision Transformers for anomaly detection in HEP. Second, we demonstrated that ensemble learning significantly improves model performance in scenarios prone to degradation, validating its effectiveness across both HEP and semiconductor datasets. These results further support ensemble learning as a powerful strategy for mitigating catastrophic forgetting and maintaining robust performance in evolving data environments.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗