Search results

Title: Large Language Model Based Machine Learning Techniques for Fake News Detection
Creator: Chen, Pin-Chien
Date: 2024
Description: With advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into...
Show moreWith advanced technology, it’s widely recognized that everyone owns one or more personal devices. Consequently, people are evolving into content creators on social media or the streaming platforms sharing their personal ideas regardless of their education or expertise level. Distinguishing fake news is becoming increasingly crucial. However, the recent research only presents comparisons of detecting fake news between one or more models across different datasets. In this work, we applied Natural Language Processing (NLP) techniques with Naïve Bayes and DistilBERT machine learning method combing and augmenting four datasets. The results show that the balanced accuracy is higher than the average in the recent studies. This suggests that our approach holds for improving fake news detection in the era of widespread content creation.
Show less

Title: Retrospective Quantitative T1 Imaging to Examine Characteristics of Multiple Sclerosis Lesions
Creator: Young, Griffin James
Date: 2024
Description: Quantitative MRI plays an essential role in assessing tissue abnormality and diseaseprogression in multiple sclerosis (MS). Specifically, T1...
Show moreQuantitative MRI plays an essential role in assessing tissue abnormality and diseaseprogression in multiple sclerosis (MS). Specifically, T1 relaxometry is gaining popularity as elevated T1 values have been shown to correlate with increased inflammation, demyelination, and gliosis. The predominant issue is that relaxometry requires parametric mapping through advanced imaging techniques not commonly included in standard clinical protocols. This leaves an information gap in large clinical datasets from which quantitative mapping could have been performed. We introduce T1-REQUIRE, a retrospective T1 mapping method that approximates T1 values from a single T1-weighted MR image. This method has already been shown to be accurate within 10% of a clinically available reference standard in healthy controls but will be further validated in MS cohorts. We also further aim to determine T1-REQUIRE’s statistical significance as a unique biomarker for the assessment of MS lesions as they relate to clinical disability and disease burden. A 14-subject comparison between T1-REQUIRE maps derived from 3D T1 weighted turbo field echoes (3D T1w TFE) and an inversion-recovery fast field echo (IRFFE) revealed a whole-brain voxel-wise Pearson’s correlation of r = 0.89 (p < 0.001) and mean bias of 3.99%. In MS white matter lesions, r = 0.81, R2 = 0.65 (p < 0.001, N = 159), bias = 10.07%, and in normal appearing white matter (NAWM), r = 0.82, R 2 = 0.67 (p < 0.001), bias = 9.48%. Mean lesional T1-REQUIRE and MTR correlated significantly (r = -0.68, p < 0.001, N = 587) similar to previously published literature. Median lesional MTR correlated significantly with EDSS (rho = -0.34, p = 0.037), and lesional T1-REQUIRE exhibited xiii significant correlations with global brain tissue atrophy as measured by brain parenchymal fraction (BPF) (r = -0.41, p = 0.010, N = 38). Multivariate linear regressions between T1- REQUIRE NAWM provided meaningful statistical relationships with EDSS (β = 0.03, p = 0.027, N = 38), as well as did mean MTR values in the Thalamus (β = -0.27, p = 0.037, N = 38). A new spoiled gradient echo variation of T1-REQUIRE was assessed as a proof of concept in a small 5-subject MS cohort compared with IR-FFE T1 maps, with a whole brain voxel-wise correlation of r = 0.88, R2 = 0.77 (p < 0.001), and Bias = 0.19%. Lesional T1 comparisons reached a correlation of r = 0.75, R2 = 0.56 (p < 0.001, N = 42), and Bias = 10.81%. The significance of these findings means that there is the potential to provide supplementary quantitative information in clinical datasets where quantitative protocols were not implemented. Large MS data repositories previously only containing structural T1 weighted images now may be used in big data relaxometric studies with the potential to lead to new findings in newly uncovered datasets. Furthermore, T1-REQUIRE has the potential for immediate use in clinics where standard T1 mapping sequences aren’t able to be readily implemented.
Show less

Title: Large-Signal Transient Stability and Control of Inverter-Based Resources
Creator: Wang, Duo
Date: 2024
Description: Renewable generation, including solar photovoltaic (PV) systems, type 3 and 4 wind turbine generation systems (WTG), battery energy storage...
Show moreRenewable generation, including solar photovoltaic (PV) systems, type 3 and 4 wind turbine generation systems (WTG), battery energy storage systems (BESS), as well as high voltage direct current (HVDC) and flexible alternating current (FACT) transmission system devices with increasing penetration level are being connected to the bulk power systems (BPS) via power electronic (PE) converters as the interface, referred to as the inverter-based resources (IBRs) on the transmission and sub-transmission levels or distributed energy resources (DERs) located on the distribution level. The IBR is almost entirely defined by the control algorithms and found to be more prone to experiencing large disturbances due to the lack of the conventional synchronous machine (SM) intrinsic synchronous characteristics and mechanical inertia, as well as being in smaller capacity sizes. Thus, these reasons motivate this dissertation to study the large-signal transient stability and control of IBRs for reliable grid integration and rapid grid transformation. For large-signal stability analysis methods, Lyapunov-based methods are the fundamental theory used to characterize the stability issues with analytical solutions, although other non-Lyapunov methods could also be very helpful. A main difficulty hindering the widespread adoption of the Lyapunov stability analysis method is the difficulty of finding the proper Lyapunov function candidate for a higher dimensional nonlinear system. The Port-Hamiltonian (PH) nonlinear control theory is explored in this dissertation as a promising theoretical framework solution addressing this challenging issue. A PH-based tracking and robust control method is proposed to facilitate the practical application of the PH framework in IBR controls. In addition, considering the typical grid-forming (GFM) IBR control with a first-order low pass filter (LPF) block is usually involved with control saturation function for protection purposes under abnormal operating conditions with anti-windup issue in practical implementation, a PH-based bounded LPF (PH-BLPF) control is proposed to incorporate this in the large-signal PH interconnection modeling framework while preserving the robust tracking Lyapunov stability with improved transient dynamic performance and stability margin.Moreover, specific real-world transient synchronization stability issues, such as the grid voltage large fault disturbance case, are studied. In addition, to meet the recent emerging IBR grid code requirements, such as the current magnitude limitation, grid support function, and fault recovery capability of GFM-VSCs, a virtual impedance-based current-limiting GFM control with enhanced transient stability and grid support is proposed.
Show less

Title: Two Essays on Mergers and Acquisitions
Creator: Xu, Yang
Date: 2024
Description: This dissertation is composed of two self-contained chapters that both relate to mergers and acquisitions (M&A). In the first essay, we...
Show moreThis dissertation is composed of two self-contained chapters that both relate to mergers and acquisitions (M&A). In the first essay, we examine the Delaware (DE) reincorporation effect on firms’ post-IPO behaviors on mergers and acquisitions. We find that firms’ DE reincorporation decisions enhance the likelihood of engaging in M&A as targets. However, as a tradeoff, DE reincorporated firms get lower takeover valuations compared to stay-at-home-state firms, and the acquisition of reincorporated firms is less likely to be successful. Our second essay aims to explore the role of the options market in price discovery for M&A. We find that the predictive power of the changes in implied volatility of the target firm stock for the takeover outcome is statistically and economically significant. The risk arbitrage portfolios incorporating filters derived from the options on stocks of the target firms generate annualized risk-adjusted abnormal returns between 2.6% and 5%, depending on the portfolio weighting method, the threshold of filters for the implied volatility change, and the asset pricing models applied for abnormal returns. The results are robust to different empirical setups and are not explained by traditional factors.
Show less

Title: Heterogeneous Workloads Study towards Large-scale Interconnect Network Simulation
Creator: Wang, Xin
Date: 2023
Description: High-bandwidth, low-latency interconnect networks play a key role in the design of modern high- performance computing (HPC) systems. The ever...
Show moreHigh-bandwidth, low-latency interconnect networks play a key role in the design of modern high- performance computing (HPC) systems. The ever-increasing need for higher bandwidth and higher message rate has driven the design of low-diameter interconnect topologies like variants of dragonfly. As these hierarchical networks become increasingly dominant, interference caused by resource sharing can lead to significant network congestion and performance variability. Meanwhile, with the rapid growth of the machine learning applications, the workloads of future HPC systems are anticipated to be a mix of scientific simulation, big data analytics, and machine learning applications. However, little work has been conducted to understand performance implications of co-running heterogeneous workloads on large-scale dragonfly systems. There is a greater need to study how different interconnect technologies affect workload performance, and how conventional scientific applications interact with emerging big data applications at the underlying interconnect level. In this work, we firstly present a comparative analysis exploring the communication interference for traditional HPC applications by analyzing the trade-off between localizing communication and balancing network traffic. We conduct trace-based simulations for applications with different communication patterns, using multiple job placement policies and routing mechanisms. Then we develop a scalable workload manager that provides an automatic framework to facilitate hybrid workload simulation. We investigate various hybrid workloads and navigate various application-system configurations for a deeper understanding of performance implications of a diverse mix of workloads on current and future supercomputers. Finally, we propose a scalable framework, Union+, that enables simulation of communication and I/O simultaneously. By combining different levels of abstraction, Union+ is able to efficiently co-model the communication and I/O traffic on HPC systems that equipped with flash-based storage. We conduct experiments with different system configurations, showing how Union+ can help system designers to assess the usefulness of future technologies in next-generation HPC machines.
Show less

Title: Predictive energy efficient control framework for connected and automated vehicles in heterogeneous traffic environments
Creator: Vellamattathil Baby, Tinu
Date: 2023
Description: Within the automotive industry, there is a significant emphasis on enhancing fuel efficiency and mobility, and reducing emissions. In this...
Show moreWithin the automotive industry, there is a significant emphasis on enhancing fuel efficiency and mobility, and reducing emissions. In this context, connected and automated vehicles (CAVs) represent a significant advancement, as they can optimize their acceleration pattern to improve their fuel efficiency. However, when CAVs coexist with human-driven vehicles (HDVs) on the road, suboptimal conditions arise, which adversely affect the performance of CAVs. This research analyzes the automation capabilities of production vehicles to identify scenarios where their performance is suboptimal, and proposes a merge-aware modification of adaptive cruise control (ACC) method for highway merging situations. The proposed algorithm addresses the issue of sudden gap and velocity changes in relation to the preceding vehicle, thereby reducing substantial braking during merging events, resulting in improved energy efficiency. This research also presents a data-driven model for predicting the velocity and position of the preceding vehicle, as well as a robust model predictive control (MPC) strategy that optimizes fuel consumption while considering prediction inaccuracies. Another focus of this research is a novel suggestion-based control framework in interactive mixed traffic environments leveraging the emerging connectivity between vehicles and with infrastructure. It is based on MPC to optimize the fuel efficiency of CAVs in heterogeneous or mixed traffic environments (i.e., including both CAVs and HDVs). In this suggestion-based control framework, the CAVs are considered to provide non-binding velocity and lane change suggestions to the HDVs to follow to improve the fuel efficiency of both the CAVs and the HDVs. To achieve this, the host CAV must devise its own fuel-efficient control solution and determine the recommendations to convey to its preceding HDV. It is assumed that the CAVs can communicate with the HDVs via Vehicle to Vehicle (V2V) communication, while the Signal Phase and Timing (SPaT) information is accessed via Vehicle-to- Infrastructure (V2I) communication. These velocity suggestions remain constant for a predefined period, allowing the driver to adjust their speed accordingly. It is also considered that the suggestions are non binding, i.e., a driver can choose not to follow the suggested velocity. For this control framework to function, we present a velocity prediction model based on experimental data that captures the response of a HDV to different suggested velocities, and a robust approach to ensure collision avoidance. The velocity prediction’s accuracy is also validated with the experimental data (on a table-top drive simulator), and the results are presented. In cases of low CAV penetration, a CAV needs to provide suggestions to multiple surrounding HDVs and incorporating the suggestions to all the HDVs as decision variables to the optimal control problem can be computationally expensive. Hence, a suggestion-based hierarchical energy efficient control framework is also proposed in which a CAV takes into account the interactive nature of the environment by jointly planning its own trajectory and evaluating the suggestions to the surrounding HDVs. Joint planning requires solving the problem in joint state- and action-space, and this research develops a Monte Carlo Tree Search (MCTS)-based trajectory planning approach for the CAV. Since the joint action- and state-space grows exponentially with the number of agents and can be computationally expensive, an adaptive action-space is proposed through pruning the action-space of each agent so that the actions resulting in unsafe trajectories are eliminated. The trajectory planning approach is followed by a low-level model predictive control (MPC)-based motion controller, which aims at tracking the reference trajectory in an optimal fashion. Simulation studies demonstrate the proposed control strategy’s efficacy compared to existing baseline methods.
Show less

Title: Integrating Deep Learning And Innovative Feature Selection For Improved Short-Term Price Prediction In Futures Markets
Creator: Tian, Tian
Date: 2024
Description: This study presents a novel approach for predicting short-term price movements in futures markets using advanced deep-learning models, namely...
Show moreThis study presents a novel approach for predicting short-term price movements in futures markets using advanced deep-learning models, namely LSTM, CNN_LSTM, and GRU_LSTM. By incorporating cophenetic correlation in feature preparation, the study addresses the challenges posed by sudden fluctuations and price spikes while maintaining diversification and utilizing a limited number of variables derived from daily public data. However, the effectiveness of adding features relies on appropriate feature selection, even when employing powerful deep-learning models. To overcome this limitation, an innovative feature selection method is proposed, which combines cophenetic correlation-based hierarchical linkage clustering with the XGBoost importance listing function. This method efficiently identifies and integrates the most relevant features, significantly improving price prediction accuracy. The empirical findings contribute valuable insights into price prediction accuracy and the potential integration of algorithmic and intuitive approaches in futures markets. Moreover, the developed feature preparation method enhances the performance of all deep learning models, including LSTM, CNN_LSTM, and GRU_LSTM. This study contributes to the advancement of price prediction techniques by demonstrating the potential of integrating deep learning models with innovative feature selection methods. Traders and investors can leverage this approach to enhance their decision-making processes and optimize trading strategies in dynamic and complex futures markets.
Show less

Title: Design and Synthesis of New Sulfur Cathodes Containing Polysulfide Adsorbing Materials
Creator: Suzanowicz, Artur M
Date: 2023
Description: Lithium-sulfur battery (LSB) technology has tremendous prospects to substitute lithium-ion battery (LIB) technology due to its high...
Show moreLithium-sulfur battery (LSB) technology has tremendous prospects to substitute lithium-ion battery (LIB) technology due to its high theoretical specific capacity and energy density. However, escaping polysulfide intermediates (produced during the redox reaction process) from the cathode structure is the primary reason for rapid capacity fading. Suppressing the polysulfide shuttle (PSS) is a viable solution for this technology to move closer to commercialization and supersede the established LIB technology. In this dissertation, I have analyzed the challenges faced by LSBs and selected methods and materials to address these problems. I have concluded that in order to further pioneer LSBs, it is necessary to address these essential features of the sulfur cathode: superior electrical conductivity to ensure faster redox reaction kinetics and high discharge capacity, high pore volume of the cathode host to maximize sulfur loading/utilization, and polar polysulfide-resistive materials to anchor and suppress the migration of lithium polysulfides.Furthermore, a versatile, low-cost, and practical scalable synthesis method is essential for translating bench-level development to large-scale production. This dissertation covers designing and synthesizing new scalable cathode structures for lithium-sulfur batteries that are inexpensive and highly functional. The rationally chosen cathode components accommodate sulfur, suppress the migration of polysulfide intermediates via chemical interactions, enhance redox kinetics, and provide electrical conductivity to sulfur, rendering excellent electrochemical performance in terms of high initial specific capacity and good long-term cycling performance. TiO2, Ni12P5, and g-C3N4 as polysulfide adsorbing materials (PAMs) have been fully studied in this thesis along with three distinct types of host structures for lithium-sulfur batteries: Polymer, Carbon Cloth, and Reduced Graphene Oxide. I have created adaptable bulk synthesis techniques that are inexpensive, easily scalable, and suitable for bench-level research as well as large-scale manufacturing. The exceptional performance and scalability of these materials make my cathodes attractive options for the commercialization of lithium-sulfur batteries.
Show less

Title: Effect of organic acid treatment in reducing Salmonella on six types of sprout seeds
Creator: Yang, Dachuan
Date: 2023
Description: Fresh sprouts present a special food safety concern as their growing conditions also favor the growth of pathogens such as Salmonella....
Show moreFresh sprouts present a special food safety concern as their growing conditions also favor the growth of pathogens such as Salmonella. Contamination in sprouts often originates from the seeds used for sprouting. The Produce Safety Rule requires that seeds used to grow sprouts be treated to reduce pathogens. The treatments may be applied by sprout growers or by seed suppliers. Although 20,000 ppm calcium hypochlorite is the most used seed treatment method, the high chlorine level can be hazardous to workers and the environment. Alternative seed treatment methods that are safe and environmentally friendly are needed. In addition, a post-treatment drying step is needed when seed suppliers are using chemical seed treatment methods. This study evaluated the efficacy of an organic acid solution for reducing Salmonella on six types of seeds (alfalfa, clover, radish, mung bean, onion, and broccoli). The impact of treatment on seed germination and sprout yield was also examined. Ten grams of seeds inoculated with a five-serotype cocktail of Salmonella were pre-rinsed with 40 ml of water twice and treated with 75.7 ml of the organic acid solution for 1 hour. The treated seeds were either not rinsed or rinsed with 40 ml of water twice before being dried in the biological safety cabinet for 24 hours. The Salmonella level, germination percentage, and sprout yield of seeds treated with water, seeds treated with the organic acid solution, seeds treated with organic acid, dried, and rinsed, and seeds treated with organic acids, dried, and not rinsed were compared. Salmonella reductions that could be achieved with this organic acid solution treatment were less than 0.5 log CFU/g without drying, 0.6-2.0 log CFU/g with drying and rinse, or 1.6-2.9 log CFU/g with drying and no rinse. Drying significantly enhanced the treatment efficacy (p < 0.05 ) on alfalfa, radish, mung bean, and onion seeds. If seeds were not rinsed after treatment, the log reductions achieved on mung bean and onion seeds were significantly higher (p < 0.05). If seeds were treated and rinsed, the germination rates of six types of seeds were not affected (p > 0.05) regardless of whether the seeds were dried or not. All treatments significantly decreased the sprout yield of clover seeds by 13% (p < 0.05 ). If seeds were not rinsed after treatment, the germination rates of clover and broccoli seeds were reduced by 7 and 9%, respectively, and the sprout yield of alfalfa seeds was reduced by 35%. Overall, the organic acid solution was ineffective when compared with 20,000 ppm calcium hypochlorite in reducing Salmonella on sprout seeds, although the drying step after treatment could improve the treatment efficacy.
Show less

Title: Utilizing Image Processing in Evaluation of Fibroblast Stimulation for Collagen Remodeling
Creator: Yoon, Shin Hae
Date: 2023
Description: This research delves into the realm of image processing as a pivotal component in the evaluation of fibroblast stimulation for collagen...
Show moreThis research delves into the realm of image processing as a pivotal component in the evaluation of fibroblast stimulation for collagen remodeling. The study focuses on unraveling the intricate synergy between electrospun silk fibroin-carbon nanotube (SF-CNT) fibers and electrical stimulation, working in harmony to enhance tissue regeneration. Building upon our previous work, we successfully engineered SF-CNT fibers through the electrospinning process, yielding highly aligned structures reminiscent of natural extracellular matrix proteins. These fibers were fortified with water stability through post-treatment with ethanol vapor, while subtle additions of carbon nanotubes (CNTs) significantly improved fiber alignment, strength, and conductivity without compromising biocompatibility. This innovative platform served as a cell culture matrix for fibroblasts harvested from pelvic organ prolapse (POP) patients, facilitating electrical stimulation that triggered a substantial increase in collagen production. In this study, we harnessed the power of various image-processing software tools, including ImageJ and Python, to analyze immunostained images of fibroblasts obtained from POP patients. Under carefully tailored electrical stimulation conditions, the stimulated cells exhibited an astonishing up to 11.97-fold increase in alpha-smooth muscle actin (α-SMA) expression, unequivocally signifying the successful activation of myofibroblasts. Additionally, in an animal model employing LOX-knockout mice to mimic collagen disorders associated with POP, the application of optimized electrical stimulation conditions for patient 003 led to a remarkable surge in collagen production and structural enhancement, underlining the potential of electrical stimulation in expediting tissue remodeling. Intriguingly, fibroblasts from patient 005 and patient 006 exhibited a distinct response, shedding light on the influence of POP severity on cellular behavior. This study firmly reinforces the imperative of personalized therapeutic approaches, emphasizing the need to customize treatment strategies to align with individual patient characteristics through innovative biological image analysis techniques.
Show less

Title: DEVELOPMENT AND APPLICATION OF A NATIONALLY REPRESENTATIVE MODEL SET TO PREDICT THE IMPACTS OF CLIMATE CHANGE ON ENERGY CONSUMPTION AND INDOOR AIR QUALITY (IAQ) IN U.S. RESIDENCES
Creator: Fazli, Torkan
Date: 2020
Description: Americans spend most of their time inside residences where they are exposed to a number of pollutants of both indoor and outdoor origin....
Show moreAmericans spend most of their time inside residences where they are exposed to a number of pollutants of both indoor and outdoor origin. Residential buildings also account for over 20% of total primary energy consumption in the U.S. and a similar proportion of greenhouse gas emissions. Moreover, climate change is expected to affect building energy use and indoor air quality (IAQ) through both building design (i.e., via our societal responses to climate change) and building operation (i.e., via changing meteorological and ambient air quality conditions). The overarching objectives of this work are to develop a set of combined building energy and indoor air mass balance models that are generally representative of both the current (i.e., ~2010s) and future (i.e., ~2050s) U.S. residential building stock and to apply them using both current and future climate scenarios to estimate the impacts of climate change and climate change policies on building energy use, IAQ, and the prevalence of chronic health hazards in U.S. homes. The developed model set includes over 4000 individual building models with detailed characteristics of both building operation and indoor pollutant physics/chemistry, and is linked to a disability-adjusted life years (DALYs) approach for estimating chronic health outcomes associated with indoor pollutant exposure. The future building stock model incorporates a combination of predicted changes in future meteorological conditions, ambient air quality, the U.S. housing stock, and population demographics. Using the model set, we estimate the total site and source energy consumption for space conditioning in U.S. residences is predicted to decrease by ~37% and ~20% by mid-century (~2050s) compared to 2012, respectively, driven by decreases in heating energy use across the building stock that are larger than coincident increases in cooling energy use in warmer climates. Indoor concentrations of most pollutants of ambient origin are expected to decrease, driven by predicted reductions in ambient concentrations due to tighter emissions controls, with one notable exception of ozone, which is expected to increase in future climate scenarios. This work provides the first known estimates of the potential magnitude of impacts of expected climate changes on building energy use, IAQ, and the prevalence of chronic health hazards in U.S. homes.
Show less

Title: Data-Driven Modeling for Advancing Near-Optimal Control of Water-Cooled Chillers
Creator: Salimian Rizi, Behzad
Date: 2023
Description: Hydronic heating and cooling systems are among the most common types of heating and cooling systems installed in older existing buildings,...
Show moreHydronic heating and cooling systems are among the most common types of heating and cooling systems installed in older existing buildings, especially commercial buildings. The results of this study based on the Commercial Building Energy Consumption Survey (CBECS) indicates chillers account for providing cooling in more than half of the commercial office building floorspaces in the U.S. Therefore, to address the need of improving energy efficiency of chillers systems operation, research studies developed different models to investigate different chiller sequencing approaches. Engineering-based models and empirical models are among the popular approaches for developing prediction models. Engineering-based models utilize the physical principles to calculate the thermal dynamics and energy behaviors of the systems and require detailed system information, while the empirical models deploy machine learning algorithms to develop relationships between input and output data. The empirical models compared to the engineering-based approach are more practical in a system’s energy prediction because of accessibility to required data, superiority in model implementation and prediction accuracy. Moreover, selecting near accurate chiller prediction models for the chiller sequencing needs to consider the importance of each input variable and its contribution to the overall performance of a chiller system, as well as the ease of application and computational time. Among the empirical modeling methods, ensemble learning techniques overcome the instability of the learning algorithm as well as improve prediction accuracy and identify input variable importance. Ensemble models combine multiple individual models, often called base or weak models, to produce a more accurate and robust predictive model. Random Forest (RF) and Extra Gradient Boosting (XGBoost) models are considered as ensemble models which offer built-in mechanisms for assessing feature importance. These techniques work by measuring how much each feature contributes to the overall predictive performance of the ensemble.In the first objective of this work the frequency of hydronic cooling systems in the U.S. building stock for applying potential energy efficiency measures (EEMs) on chiller plants are explored. Results show that the central chillers inside the buildings are responsible for providing cooling for more than 50% of the commercial buildings with areas greater than 9,000 m2(~100,000 ft2). In addition, hydronic cooling systems contribute to the highest Energy Use Intensity (EUI) among other systems, with EUI of 410.0 kWh/m2 (130.0 kBtu/ft2). Therefore, the results of this objective support developing accurate prediction models to assess the chiller performance parameters as an implication for chiller sequencing control strategies in older existing buildings. The second objective of the dissertation is to evaluate the performance of chiller sequencing strategy for the existing water-cooled chiller plant in a high-rise commercial building and develop highly accurate RF chiller models to investigate and determine the input variables of greatest importance to chiller power consumption predictions. The results show that the average value of mean absolute percentage error (MAPE) and root mean squared error (RMSE) for all three RF chiller models are 5.3% and 30 kW, respectively, for the validation dataset, which confirms a good agreement between measured and predicted values. On the other hand, understanding prediction uncertainty is an important task to confidently reporting smaller savings estimates for different chiller sequencing control strategies. This study aims to quantify prediction uncertainty as a percentile for selecting an appropriate confidence level for chillers models which could lead to better prediction of the peak electricity load and participate in demand response programs more efficiently. The results show that by increasing the confidence level from 80% to 90%, the upper and lower bounds of the demand charge differ from the actual value by a factor of 3.3 and 1.7 times greater, respectively. Therefore, it proves the significance of selecting appropriate confidence levels for implementation of chiller sequencing strategy and demand response programs in commercial buildings. As the third objective of this study, the accuracy of these prediction models with respect to the preprocessing, selection of data, noise analysis, effect of chiller control system performance on the recorded data were investigated. Therefore, this study attempts to investigate the impacts of different data resolution, level of noise and data smoothing methods on the chiller power consumption and chiller COP prediction based on time-series Extra Gradient Boosting (XGBoost) models. The results of applying the smoothing methods indicate that the performance of chiller COP and the chiller power consumption models have improved by 2.8% and 4.8%, respectively. Overall, this study would guide the development of data-driven chiller power consumption and chiller COP prediction models in practice.
Show less

Title: Scalable Indexing and Search in High-End Computing Systems
Creator: Orhean, Alexandru Iulian
Date: 2023
Description: Rapid advances in digital sensors, networks, storage, and computation coupled with decreasing costs is leading to the creation of huge...
Show moreRapid advances in digital sensors, networks, storage, and computation coupled with decreasing costs is leading to the creation of huge collections of data. Increasing data volumes, particularly in science and engineering, has resulted in the widespread adoption of parallel and distributed file systems for storing and accessing data efficiently. However, as file system sizes and the amount of data ``owned” by users grows, it is increasingly difficult to discover and locate data amongst the petabytes of data. While much research effort has focused on methods to efficiently store and process data, there has been relatively little focus on methods to efficiently explore, index, and search data using the same high-performance storage and compute systems. Users of large file systems either invest significant resources to implement specialized data catalogs for accessing and searching data, or resort to software tools that were not designed to exploit modern hardware. While it is now trivial to quickly discover websites from the billions of websites accessible on the Internet, it remains surprisingly difficult for users to search for data on large-scale storage systems. We initially explored the prospect of using existing search engine building blocks (e.g. CLucene) to integrate search in a high-performance distributed file system (e.g. FusionFS), by proposing and building the FusionDex system, a distributed indexing and query model for unstructured data. We found indexing performance to be orders of magnitude slower than theoretical speeds we could achieve in raw storage input and output, and sought to investigate a new clean-slate design for high-performance indexing and search.We proposed the SCANNS indexing framework to address the problem of efficiently indexing data in high-end systems, characterized by many-core architectures, with multiple NUMA nodes and multiple PCIe NVMe storage devices. We designed SCANNS as a single-node framework that can be used as a building block for implementing high-performance indexed search engines, where the software architecture of the framework is scalable by design. The indexing pipeline is exposed and allows easy modification and tuning, enabling SCANNS to saturate storage, memory and compute resources on different hardware. The proposed indexing framework uses a novel tokenizer and inverted index design to achieve high performance improvement both in terms of indexing and in terms of search latency. Given the large amounts and the variety of data found in scientific large-scale file systems, it stands to reason to try to bridge the gap between various data representations and to build and provide a more uniform search space. ScienceSearch is a search infrastructure for scientific data that uses machine learning to automate the creation of metadata tags from different data sources, such as published papers, proposals, images and file system structure. ScienceSearch is a production system that is deployed on a container service platform at NERSC and provides search over data obtained from NCEM. We conducted a performance evaluation of the ScienceSearch infrastructure focusing on scalability trends in order to better understand the implications of performing search over an index built from the generated tags. Drawing from the insights gained from SCANNS and the performance evaluation of ScienceSearch, we explored the problems of efficiently building and searching persistent indexes that do not fit into main memory. The SCIPIS framework builds on top of SCANNS and further optimizes the inverted index design and indexing pipeline, by exposing new tuning parameters that allows the user to further adapt the index to the characteristics of the input data. The proposed framework allows the user to quickly build a persistent index and to efficiently run TFIDF queries over the built index. We evaluated SCIPIS over three kinds of datasets (logs, scientific data, and file system metadata) and showed that it achieves high indexing and search performance and good scalability across all datasets.
Show less

Title: Colored Pencil Drawings, undated
Creator: Henry, Mary Dill, 1913-2009
Description: Untitled colored pencil drawings by Mary Henry, date unknown. Inscription on verso: "William Winter Comments, PO Box 817, Sausalito"
Collection: Mary Dill Henry Papers, 1913-2021

Title: Computational Genomics of Human-Infecting Microsporidia Species from the Genus Encephalitozoon
Creator: Mascarenhas dos Santos, Anne Caroline
Date: 2023
Description: Microsporidia are obligate intracellular pathogens classified as category B priority pathogens by the National Institute of Allergy and...
Show moreMicrosporidia are obligate intracellular pathogens classified as category B priority pathogens by the National Institute of Allergy and Infectious Diseases (NIAID), a division of the National Institutes of Health (NIH). Microsporidian species from the genus Encephalitozoon infect humans and can cause encephalitis, keratoconjunctivitis or enteric diseases in both immunocompromised and immunocompetent individuals. The main treatment for disseminated microsporidiosis available in the United States is albendazole, an anthelmintic benzimidazole that is also used to treat fungal infections, but species from the Encephalitozoonidae have already shown signs of resistance against this drug. The Encephalitozoonidae harbors highly specialized pathogens with the smallest known eukaryote genomes, with Encephalitozoon cuniculi featuring a genome of only 2.9 Mbp and coding for a proteome of roughly 2,000 proteins. Pathogens are in an everlasting race to quicken their adaptation pace against host defenses. This adaptation is often driven by gene duplication, recombination and/or mutation, and due to the potentially disruptive nature of duplication and recombination processes, many of these evolutions in pathogens are taking place outside conserved genomic loci. As such, genes involved in virulence and drug resistance in pathogens are often localized in the (sub)telomeres rather than in chromosome cores. The small and streamlined nature of microsporidian genomes makes them excellent candidates to investigate the adaptation of pathogens to host defenses, the evolution of their virulence, and the development of their resistance to drugs from a genomic perspective. However, microsporidian genomes are highly divergent at the DNA sequence level and the ones that have been sequenced so far are incomplete and are lacking the telomeres. This high level of sequence divergence hinders standard sequence homology-based functional annotations, blurring our understanding of what these organisms are capable of from a metabolic perspective. The gap in our knowledge of what is encoded in the microsporidia telomeres could lead to an underestimation of their pathogenic capabilities. Therefore, deciphering the functions of unknown proteins in microsporidia genomes and unraveling the content of their telomeres is important to fully assess their potential for adaptability to host defenses and predisposition to drug resistance. Likewise, a better understanding of the genetic diversity in microsporidia will help assess the extent by which host-pathogen interactions are shaping the adaptation of these parasites to humans. As observed in the COVID-19 pandemic, genetic diversity can influence the speed at which pathogens adapt to host defenses and thus can pose a big challenge to disease control. The development of strategies for controlling microsporidiosis outbreaks will likely benefit from the work performed in this thesis. As part of my PhD work, I investigated the virulence and host-adaptation capabilities of human-infecting microsporidia species from the genus Encephalitozoon with computational genomic approaches. This work included: 1) using structural homology to infer the functions of unknown proteins from the microsporidia proteome; 2) sequencing the complete genomes from telomere-to-telomere of three distinct Encephalitozoon spp. (E. cuniculi, E. hellem and E. intestinalis) to determine the genetic makeup of their telomeres and better understand the extent of their diversity; and 3) assessing the intraspecies genetic diversity that exists between Encephalitozoon species.
Show less

Title: Eating disorder support group utilization: Associations with psychological health and eating disorder psychopathology among support group attendees
Creator: Murray, Matthew F.
Date: 2023
Description: Individuals with eating disorders (EDs) report psychosocial impairments that may persist beyond ED symptom remission, suggesting a need to...
Show moreIndividuals with eating disorders (EDs) report psychosocial impairments that may persist beyond ED symptom remission, suggesting a need to examine ED treatment-adjunctive services that foster psychosocial health. One promising resource is support groups, as evidence across medical and psychiatric illnesses shows associations between group utilization and wellbeing. However, virtually no literature has examined ED-specific support groups and psychosocial health, and it is also unknown how use of supportive services relates to ED symptoms. The present study examined associations between past-month ED support group attendance and participation frequency, psychosocial health indices, and ED symptoms. A total of 215 participants who attended weekly virtual clinician-moderated ED support groups completed measures of psychosocial health, internalized stigma of mental illness, psychosocial impairment from an ED, specific types of social support elicited in group, and ED psychopathology. Adjusting for past-month ED treatment, Benjamini-Hochberg-corrected partial correlation analyses indicated that more frequent attendance was negatively related to body dissatisfaction, purging, excessive exercise, and negative attitudes toward obesity, and positively related to social support. More frequent verbal and chat participation were positively related to emotional and informational support and social companionship. Chat participation was additionally negatively related to excessive exercise and negative attitudes toward obesity. Results suggest that utilizing and participating in clinician-moderated ED support groups could provide an outlet for ED symptom management and solicitation of social support. Findings highlight areas for further consideration in the delivery of and future research on ED support groups.
Show less

Title: Optimization methods and machine learning model for improved projection of energy market dynamics
Creator: Saafi, Mohamed Ali
Date: 2023
Description: Since signing the legally binding Paris agreement, governments have been striving to fulfill the decarbonization mission. To reduce carbon...
Show moreSince signing the legally binding Paris agreement, governments have been striving to fulfill the decarbonization mission. To reduce carbon emissions from the transportation sector, countries around the world have created a well-defined new energy vehicle development strategy that is further expanding into hydrogen vehicle technologies. In this study, we develop the Transportation Energy Analysis Model (TEAM) to investigate the impact of the CO2 emissions policies on the future of the automotive industries. On the demand side, TEAM models the consumer choice considering the impacts of technology cost, energy cost, refueling/charging availability, consumer travel pattern. On the supply side, the module simulates the technology supply by the auto-industry with the objective of maximizing industry profit under the constraints of government policies. Therefore, we apply different optimization methods to guarantee reaching the optimal automotive industry response each year up to 2050. From developing an upgraded differential evolution algorithm, to applying response surface methodology to simply the objective function, the goal is to enhance the optimization performance and efficiency compared to adopting the standard genetic algorithm. Moreover, we investigate TEAM’s robustness by applying a sensitivity analysis to find the key parameters of the model. Finally based on the key sensitive parameters that drive the automotive industry, we develop a neural network to learn the market penetration model and predict the market shares in a competitive time by bypassing the total cost of ownership analysis and profit optimization. The central motivating hypothesis of this thesis is that modern optimization and modeling methods can be applied to obtain a computationally-efficient, industry-relevant model to predict optimal market sales shares for light-duty vehicle technologies. In fact, developing a robust market penetration model that is optimized using sophisticated methods is a crucial tool to automotive companies, as it quantifies consumer’s behavior and delivers the optimal way to maximize their profits by highlighting the vehicles technologies that they could invest in. In this work, we prove that TEAM reaches the global solution to optimize not only the industry profits but also the alternative fuels optimized blends such as synthetic fuels. The time complexity of the model has been substantially improved to decrease from hours using the genetic algorithm, to minutes using differential evolution, to milliseconds using neural network.
Show less

Title: Migration of Silver from Silver Zeolite/Low-Density Polyethylene Films into Food Stimulants
Creator: Sayeed, Maryam
Date: 2023
Description: Zeolites are naturally occurring or synthetic crystalline microporous aluminosilicate structures with remarkable catalytic, adsorption, and...
Show moreZeolites are naturally occurring or synthetic crystalline microporous aluminosilicate structures with remarkable catalytic, adsorption, and ion-exchange properties. Their unique framework of pores, channels, and cages with precise dimensions makes them an excellent fit for ion exchange and storage. Silver-exchanged zeolite (Ag/Y) composites may be incorporated into polymer matrices to create antimicrobial packaging materials. The slow release of Ag from nanosilver-enabled polymer nanocomposites (PNCs) may inhibit the growth of bacteria and other pathogens on the film’s surface, improving food quality and reducing food waste. However, the migration of Ag ions from the film into food matrices is of great concern as it could expose humans to high concentrations of a heavy metal from dietary sources. The amount of migration depends on various factors, including the potential form of Ag and its concentration in the film, the film thickness, and the storage conditions.The primary objective of this study is to investigate the effect of the form of Ag bound to the zeolite on the migration behavior of Ag from Ag/Y incorporated low-density polyethylene (LDPE) films. For Ag/Y-incorporated LDPE PNCs with distinct Ag species, the Ag migration into the water and Squirt (a commercial soft drink) was at least four times higher from films containing zeolites exchanged with ionic Ag versus zeolites exchanged with nanoparticulate Ag. Similarly, migration into 9 wt % aqueous Domino sugar (granulated sucrose) solution was seven times higher in the ionic silver-incorporated film than in the nanoparticulate Ag film. This study suggests that it is important to consider the form of Ag in silver-exchanged zeolite while producing packaging materials since the potential form of Ag in the PNCs might significantly affect Ag migration behavior.
Show less

Title: Evaluating Speech Separation Through Pre-Trained Deep Neural Network Models
Creator: Prabhakar, Deeksha
Date: 2023
Description: Speaker separation involves separating individual speakers from a mixture of voices or background noise, known as the "cocktail party problem....
Show moreSpeaker separation involves separating individual speakers from a mixture of voices or background noise, known as the "cocktail party problem." This refers to the ability to focus on a specific sound while filtering out other distractions.In this analysis, we propose the idea of obtaining features present in the original data and then evaluating the impact they have on the ability of the model to separate the mixed audio streams. The dataset is prepared such that these feature values can be used as predictor variables to various models like Logistic Regression, Decision Trees, SVM (both rbf and linear kernel), XGBoost, AdaBoost, to obtain the most contributing features that is the features that will lead to a better separation. These results shall then be analyzed to conclude the features that affect separating the audio streams the most. Initially, 400 audio streams are selected from the VoxCeleb dataset and combined to form 200 single utterances. After the mixes are obtained, the pre-trained Speechbrain model, sepformer-whamr is used. This model separates the audio mixes given as input and obtain two outputs that should be as close as possible to the original ones. A feature list from the 400 chosen audios is obtained and then the effect of certain features on the model's capability to distinguish between multiple audio sources in a mixed recording is assessed. Two analysis parameters- permutation feature importance and SHAP values are used to conclude which features have more effect on separation. Our hypothesis is that the features contributing the most to a good separation are invariant across datasets. To test this hypothesis, we obtain 1,000 audio streams from the Mozilla Common Voice Dataset and perform the same experimental methodology described above. Our results demonstrate that the features we extract from VoxCeleb dataset are indeed invariant and aid in separating the audio streams of the Mozilla Common Voice dataset.
Show less

Title: Improving self-supervised monocular depth estimation from videos using forward and backward consistency
Creator: Shen, Hui
Date: 2020
Description: Recently, there has been a rapid development in monocular depth estimation based on self-supervised learning. However, these existing self...
Show moreRecently, there has been a rapid development in monocular depth estimation based on self-supervised learning. However, these existing self-supervised learning methods are insufficient for estimating motion objects, occlusions, and large static areas. Uncertainty or vanishing easily occurs during depth inferencing. To address this problem, the model proposed in this thesis further explores the consistency in video and builds a multi-frame model for depth estimation; secondly, by taking advantage of the optical flow, a motion mask is generated, with additional photometric loss applied for those masked regions. Experiments are carried out on the KITTI dataset. The proposed model performs better than the baseline model in quantitative results, and as seen from the depth map, the scale uncertainty and depth incomplete situations are improved in motion objects and occlusions explicitly.
Show less

repository.iit

Search the repository

Pages

Pages

Enabled Filters

Refine Results

Type

Date

Academic status of creator

Department

Subject

Creator

Part of

Rights