Engineering Commons^™

Scene Classification Of Remote Sensing Image Based On Multi-Path Reconfigurable Neural Network, Wenyi Hu, Chunjie Lan, Tian Chen, Shan Liu, Lirong Yin, Lei Wang

Electrical & Computer Engineering Faculty Publications

Land image recognition and classification and land environment detection are important research fields in remote sensing applications. Because of the diversity and complexity of different tasks of land environment recognition and classification, it is difficult for researchers to use a single model to achieve the best performance in scene classification of multiple remote sensing land images. Therefore, to determine which model is the best for the current recognition classification tasks, it is often necessary to select and experiment with many different models. However, finding the optimal model is accompanied by an increase in trial-and-error costs and is a waste of …

Investigation Of The Effect Of Preparation Parameters On The Structural And Mechanical Properties Of Gelatin/Elastin/Sodium Hyaluronate Scaffolds Fabricated By The Combined Foaming And Freeze-Drying Techniques, Mansour Qamash, S. Misagh Imani, Meisam Omidi, Ciara Glancy, Lobat Tayebi

Electrical & Computer Engineering Faculty Publications

This paper aimed to evaluate the effects of different preparation parameters, including agitation speed, agitation time, and chilling temperature, on the structural and mechanical properties of a novel gelatin/elastin/sodium hyaluronate tissue engineering scaffold, recently developed by our research group. Fabricated using a combination of foaming and freeze-drying techniques, the scaffolds were assessed to understand how these parameters influence their morphology, internal microstructure, porosity, mechanical properties, and degradation behavior. The fabrication process used in this study involved preparing a homogeneous aqueous solution containing 8% gelatin, 2% elastin, and 0.5% sodium hyaluronate (w/v), which was then subjected to mechanical agitation at speeds …

Generalized Functions In The Study Of Signals And Systems, Erik I. Verriest, Gunther Dirr, W. Steven Gray

Electrical & Computer Engineering Faculty Publications

We collect three instances where the theory of generalized functions may still make contributions to the study of signals and systems. In the first, a purely algebraic approach is presented for LTI-ODE's, in terms of two operators, D and T, respectively the differentiation operator and the multiplication-by-the-independent-variable operator. This formalism adds simplicity, a duality theory, and nicely generalizes to other classes of operator equations and their solutions. In the second part we extend the classical bilateral Laplace transform to include Bohl functions with support in ℝ by invoking Sato's hyperfunctions. Finally, in the third case we use the Colombeau algebra …

An Enhanced Real-Time Intrusion Detection Framework Using Federated Transfer Learning In Large-Scale Iot Networks, Khawlah Harahsheh, Malek Alzaqebah, Chung-Hao Chen

Electrical & Computer Engineering Faculty Publications

The exponential growth of Internet of Things (IoT) devices has introduced critical security challenges, particularly in scalability, privacy, and resource constraints. Traditional centralized intrusion detection systems (IDS) struggle to address these issues effectively. To overcome these limitations, this study proposes a novel Federated Transfer Learning (FTL)-based intrusion detection framework tailored for large-scale IoT networks. By integrating Federated Learning (FL) with Transfer Learning (TL), the framework enhances detection capabilities while ensuring data privacy and reducing communication overhead. The hybrid model incorporates convolutional neural networks (CNNs), bidirectional gated recurrent units (BiGRUs), attention mechanisms, and ensemble learning. To address the class imbalance, Synthetic …

A Review Of Emerging Sensor Technologies For Tank Inspection: A Focus On Lidar And Hyperspectral Imaging And Their Automation And Deployment, Sergio Pallas Enguita, Chung-Hao Chen, Samuel Kovacic

Electrical & Computer Engineering Faculty Publications

This paper reviews various sensor technologies for tank inspection, focusing on Light Detection and Ranging (LiDAR) and Hyperspectral Imaging (HSI) as advanced solutions for corrosion detection. These technologies are evaluated alongside traditional methods such as ultrasonic, electromagnetic, and thermographic inspections. This review highlights their potential to enhance inspection accuracy, reduce the limitations of manual inspection, and support integrated data analysis for comprehensive asset management. Additionally, this paper proposes a pathway for automating these techniques to streamline inspection processes and improve implementation in practical applications.

Advancing Chronic Kidney Disease Prediction Through Machine Learning And Deep Learning With Feature Analysis, Shiddarth Dey Tusar, S. M. Ahad Ali Chowdhury, Md. Jalal Uddin Chowdhury, Rana M. Pir, H. M. Nur A. Alam, Muhammad Rezaur Rahman, Md. Nural Absar Siddiky, Muhammad Enayetur Rahman

Electrical & Computer Engineering Faculty Publications

Chronic Kidney Diesease (CKD) is a significant health issue, ranking as the fourth leading cause of mortality worldwide. The traditional diagnosis and treatment process, reliant on medical experts, is time-consuming. Therefore, thereis an urgent need for more efficient diagnostic methods to improve patient outcomes and reduce mortality rates. In this study, we employ Machine Learning (ML) and Deep Learning (DL) techniques to predict CKD based on important features. Feature analysis was performed using a correlation matrix and the LASSO algo-rithm to identify the most relevant features for model training. We evaluated several ML and DL classifiers, including Logistic Regression (LR), …

Comparative Analysis Of Machine Learning Models For Predicting Healthcare Traffic: Insights For Optimized Emergency Response, Shadman Mahmood Khan Pathan, Sakan Binte Imran, M. M. Shabab Iqbal, Muhammad Enayetur Rahman, Md. Nurul Absar Siddiky, Muhammad Rezaur Rahman, Md Rafid Hasan, Nondon Lal Dey, Md Sobuj Hossain

Electrical & Computer Engineering Faculty Publications

Efficient management of healthcare traffic is crucial for ensuring timely access to medical services, particularly in emergency situations where delays can have severe consequences. This study presents a comparative analysis of three widely used machine learning models—Linear Regression, Decision Trees, and Random Forests—aimed at predicting healthcare-related traffic volumes. A large dataset from a metropolitan traffic system was used to train and evaluate the models based on key performance indicators, including Mean Squared Error (MSE), R² Score, and computational efficiency. The results reveal that the Random Forest model offers the best performance, achieving higher predictive accuracy and faster execution times compared …

Direct Measurement Of Microwave Loss In Nb Films For Superconducting Qubits, B. Abdisatarov, D. Bafia, A. Murthy, G. Eremeev, H. E. Elsayed-Ali, J. Lee, A. Netepenko, C. P. A. Carlos, S. Leith, G. J. Rosaz, A. Romanenko, A. Grassellino

Electrical & Computer Engineering Faculty Publications

Niobium films are a key component in modern two-dimensional superconducting qubits, yet their contribution to the total qubit decay rate is not fully understood. The presence of different layers of materials and interfaces makes it difficult to identify the dominant loss channels in present two-dimensional qubit designs. In this paper, we present the study that directly correlates measurements of RF losses in such films to material parameters by investigating a high-power impulse magnetron sputtered (HiPIMS) film atop a three-dimensional niobium superconducting radio frequency (SRF) resonator. By using a 3D SRF structure, we are able to isolate the niobium film loss …

Distance Estimation Based On Step Frequency Using Accelerometer Data, Rami M. Al-Naimat, Khawlah M. Harahsheh, Chung-Hao Chen

Electrical & Computer Engineering Faculty Publications

In recent years, smartphone sensors have become one of the most important and easily available sensors to facilitate people's lives, especially in health care and positioning (indoor environments). However, the data coming from smartphone sensors can be distorted during the user’s movement such as irrelevant movements, walk mode, and speed of walking. This distortion (noise) impairs the estimated distance accuracy (accumulative error) which increases with increasing walking distance. In addition, the accuracy of the distance traveled is affected by the user's speed, as the speed affects the step length. This work proposes a novel approach for calculating step length in …

Spinel Oxide Enables High-Temperature Self-Lubrication In Superalloys, Zhengyu Zhang, Eitan Hershkovitz, Qi An, Liping Liu, Xiaqing Wang, Zhifei Deng, Garrett Baucom, Wenbo Wang, Jing Zhao, Ziming Xin, Lowell Moore, Yingqi Yao, Md Rezwan Ul Islam, Xin Chen, Bai Cui, Ling Li, Hongliang Xin, Lin Li, Honggyu Kim, Wenjun Cai

Research, Publications & Creative Work

The ability to lubricate and resist wear at temperatures above 600 °C in an oxidative environment remains a significant challenge for metals due to their high-temperature softening, oxidation, and rapid degradation of traditional solid lubricants. Herein, we demonstrate that high-temperature lubricity can be achieved with coefficients of friction (COF) as low as 0.10-0.32 at 600- 900 °C by tailoring surface oxidation in additively-manufactured Inconel superalloy. By integrating high-temperature tribological testing, advanced materials characterization, and computations, we show that the formation of spinel-based oxide layers on superalloy promotes sustained self-lubrication due to their lower shear strength and more negative formation and …

Removal Of Pfoa/Pfos In Water Using Electrocoagulation And Associated Life Cycle Impacts And Financial Costs, Tianhong Mu

Electronic Theses & Dissertations (2024 - present)

Electrocoagulation (EC) with a zinc anode demonstrated promising results to remove perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) from an aqueous solution. However, the energy requirement for EC is usually very high due to water electrolysis or aeration. The second chapter of this dissertation aims to reduce energy consumption using an air-cathode in EC (ACEC) to supply oxygen electron acceptor without aeration for attenuating PFOA/PFOS in this new configuration. For the high PFOA concentration (0.25 mM), ACEC with 45 min of the reaction time exhibited an excellent PFOA removal (99.8 ± 0.3% removal) comparable to an EC with aeration (EC-aeration, …

Aerosol Jet Printing Of Surface Acoustic Wave Microfluidic Devices, Joseph Rich, Brian Cole, Teng Li, Brandon Lu, Hanyu Fu, Brittany N. Smith, Jianping Xia, Shujie Yang, Ruoyu Zhong, James L. Doherty, Kanji Kaneko, Hiroaki Suzuki, Zhenhua Tian, Aaron D. Franklin, Tony Jun Huang

Faculty Publications

The addition of surface acoustic wave (SAW) technologies to microfluidics has greatly advanced lab-on-a-chip applications due to their unique and powerful attributes, including high-precision manipulation, versatility, integrability, biocompatibility, contactless nature, and rapid actuation. However, the development of SAW microfluidic devices is limited by complex and time-consuming micro/nanofabrication techniques and access to cleanroom facilities for multistep photolithography and vacuum-based processing. To simplify the fabrication of SAW microfluidic devices with customizable dimensions and functions, we utilized the additive manufacturing technique of aerosol jet printing. We successfully fabricated customized SAW microfluidic devices of varying materials, including silver nanowires, graphene, and poly(3,4-ethylenedioxythiophene) polystyrene sulfonate …

Leveraging Compliance To Design A Minimally Invasive, Expandable Interbody Cage Capable Of Customized Anatomical Fit For Spinal Fusion Surgery, Larry L. Howell, Daniel J. Orr, Anton E. Bowden, Christian Payne, Hailey Jones, Bruce M. Frankel, James Anderson, Alek Sperry, Brandon Sargent

Faculty Publications

As spinal fusion surgery continues to transition to less invasive techniques, there remains an unmet need for ever smaller and more complex interbody cages to meet the unique needs of this difficult surgery. This work focuses on the hypothesis that this need can be met using the inherent advantages of compliant mechanisms in a way no previous device has. Deployable Euler Spiral Connectors optimized using a gradient based optimization algorithm were used as the foundation for a device that can stow to a very small size for device insertion then bilaterally deploy to a substantially larger device footprint. Additionally, a …

Methodology For Designing And Improving Novel Kirigami Patterns Using The Hamiltonian Circuit Framework, Larry L. Howell, Nathan Coleman, Mckaelin Edralin, Specer P. Magleby, Denise Halverson, Zhong You

Faculty Publications

The ability to compactly fold arrays to be deployed in space has seen an increasing demand in the aerospace industry. In this work, we build on existing methods to create kirigami patterns which can stow compactly and deploy out to a predictable shape using Hamiltonian circuits. We detail how designers can use this method to create novel kirigami patterns for a desired deployed shape and area. We also discuss techniques for modifying and improving the behavior and performance of kirigami patterns created using this method, and then show example applications including as a deployable radio frequency antenna of this method …

Ifn-𝜸 And Il-10 Immunosensor With Vertically Aligned Carbon Nanotube Interdigitated Electrodes Toward Pen-Side Cattle Paratuberculosis Monitoring, Shaowei Ding, Benjamin J. Brownlee, Kshama Parate, Cicero C. Pola, Bolin Chen, Jesse M. Hostetter, Douglas Jones, John Jackman, Brian D. Iverson, Jonathan C. Claussen

Faculty Publications

Highly sensitive vertically aligned carbon nanotube arrays (VANTAs) interdigitated electrode (IDE) arrays are developed for electrochemical biosensing of two cytokines (i.e., interleukin-10 (IL-10) and interferon-gamma (IFN-𝜸)) that are useful for early detection Johne’s disease (Bovine Paratuberculosis) in cattle. The high aspect ratio VANTA-IDEs (50–60 μm in height) are grown through a chemical vapor deposition process from an iron (Fe) catalyst that is lithographically patterned on a silicon wafer with equal finger width and inter-finger spacing of 25 μm. After functionalization with distinct antibodies the VANTA-IDEs are capable of selective detection of both IL-10 and IFN-𝜸 within an actual biological matrix …

Review Of The Cavity Effect: Modeling And Impact Of Cavity Shape On Apparent Radiative Surface Properties, Ehsan Mofidipour, Matthew R. Jones, Brian D. Iverson

Faculty Publications

Radiative surface properties play a critical role in the analysis, design, and optimization of thermal systems. Geometry has a strong influence on the emission and absorption characteristics of a surface. Modification of radiative surface properties may be achieved by using engineered surfaces or cavities by capitalizing on the cavity effect. A model for the apparent emissivity of an arbitrarily shaped cavity with spectrally and directionally dependent radiative surface properties is presented in this work. This model is verified through comparison with published models of the apparent radiative properties of cylindrical cavities. The merit of this model is demonstrated by obtaining …

A Charging Strategy For Large Commercial Electric Vehicle Fleets, Bassam Al-Hanahi, Iftekhar Ahmad, Daryoush Habibi, Pravakar Pradhan, Mohammad A. S. Masoum

Research outputs 2022 to 2026

The popularity of Commercial Electric Vehicles (CEVs) has experienced a surge in recent years, particularly in urban vocational contexts, as a means of advancing towards the goal of attaining net-zero emissions by 2050. The return-to-base charging strategy, which involves charging CEVs at depots, has become a prevalent practice for smaller CEV fleets. Nevertheless, for larger CEV fleets, the limited charging capacity at depots presents a significant challenge, leading to a reliance on both limited depot charging infrastructure and public charging infrastructure. This reliance can have a substantial impact on both the operational costs and the sustainability of logistics services. To …

Salmonella Detection In Food Using A Hek-Htlr5 Reported Cell-Based Sensor, Esma Eser, Victoria A. Felton, Rishi Drolia, Arun K. Bhunia

Biological Sciences Faculty Publications

The development of a rapid, sensitive, specific method for detecting foodborne pathogens is paramount for supplying safe food to enhance public health safety. Despite the significant improvement in pathogen detection methods, key issues are still associated with rapid methods, such as distinguishing living cells from dead, the pathogenic potential or health risk of the analyte at the time of consumption, the detection limit, and the sample-to-result. Mammalian cell-based assays analyze pathogens’ interaction with host cells and are responsive only to live pathogens or active toxins. In this study, a human embryonic kidney (HEK293) cell line expressing Toll-Like Receptor 5 (TLR-5) …

Receptor-Targeted Next-Generation Probiotics Ameliorate Mammalian Colitis, Nicholas L. F. Gallina, Vignesh Nathan, Akshay Krishnakumar, Dongqi Liu, Rishi Drolia, Nicole Irrizary Tardi, Yang Fu, Manalee Samadar, Shivendra Tenguria, Alvin Cai, Ruth Eunice Centeno-Martinez, Timothy A. Johnson, Abigail Cox, Lavanya Reddivari, Bruce Applegate, Xingjian Bai, Luping Xu, Deepti Tanjore, Ramesh Vemulapalli, Rahim Rahimi, Arun K. Bhunia

Biological Sciences Faculty Publications

Introduction: a loss of intestinal barrier function, inflammation, and an elevated expression of epithelial heat shock protein 60 (Hsp60) are features of an inflamed bowel. Probiotics have been used to alleviate colitis-induced pathologies, but offer poor adhesion and adaptation to the diseased gut. We hypothesize that enhancing probiotic adhesion in the inflamed bowel may ameliorate such pathologies. Listeria adhesion protein (LAP; 94-kDa acetaldehyde alcohol dehydrogenase) aids Listeria attachment to the epithelial cells by interacting with the mammalian receptor Hsp60. Bioengineered Lactobacillus casei probiotics (BLPs) expressing LAP showed strong interaction with epithelial Hsp60, a high immunomodulatory response, and sustained epithelial barrier …

Compare And Contrast The Intent Of Hacking Vs Penetration Testing, Kehinde Alabi

Harrisburg University Other Works

No abstract provided.

Contrast And Compare The Cyber Hacking Laws Between The United States, Russian Federation, And The People's Republic Of China, Kehinde Alabi

Harrisburg University Other Works

Cyber hacking is a growing threat in the modern world. As the world becomes more digital, the threat of cybercrime continues to grow. Cyberattacks can lead to stolen personal information, financial losses and damage to critical information. The increase in the digital transformation of companies has also led to an increase in cyber security concerns. With cybercriminals using increasingly advanced methods to gain access and take advantage of sensitive information. As a result, governments around the world have developed measures, laws, and regulations to address cybercrime and protect against cyberattacks. The United States, Russian Federation, and the People’s Republic of …

การจำแนกและการประเมินประสบการณ์ผู้ใช้จากบทวิจารณ์ของผู้ใช้โมไบล์แอปพลิเคชันโดยใช้การเรียนรู้ของเครื่อง, จิราพร รามจุล

Chulalongkorn University Theses and Dissertations (Chula ETD)

การทำความเข้าใจประสบการณ์ของผู้ใช้ เป็นสิ่งสำคัญต่อความสำเร็จและการพัฒนาอย่างต่อเนื่องของโมไบล์แอปพลิเคชัน งานวิจัยนี้นำเสนอวิธีการจำแนกบทวิจารณ์ของผู้ใช้ตามลักษณะของประสบการณ์ผู้ใช้ที่มีต่อโมไบล์แอปพลิเคชัน โดยใช้ทั้งอัลกอริทึมการเรียนรู้ของเครื่อง และการเรียนรู้เชิงลึก บทวิจารณ์ของผู้ใช้ถูกจำแนกตามมิติของประสบการณ์ผู้ใช้ในด้านการปฏิบัติและประสบการณ์ผู้ใช้ในด้านความเพลิดเพลิน ซึ่งประกอบด้วยลักษณะของประสบการณ์ผู้ใช้ทั้งหมด 8 รายการ ตามที่นิยามไว้ในแบบสอบถามประสบการณ์ผู้ใช้แบบสั้น ผลลัพธ์การจำแนกบทวิจารณ์ของผู้ใช้สามารถนำไปใช้ร่วมกับการวิเคราะห์ความรู้สึกจากบทวิจารณ์ เพื่อประเมินคะแนนประสบการณ์ผู้ใช้โดยรวมของโมไบล์แอปพลิเคชันได้อย่างอัตโนมัติ จากการทดลองประเมินประสิทธิภาพของโมเดลพบว่าเบิร์ต มีประสิทธิภาพดีกว่าซัพพอร์ตเวกเตอร์แมชีน แรนดอมฟอเรสต์ และ ลอจิสติกรีเกรสชัน โดยมีค่าความเที่ยง เท่ากับ 80%, ค่าเรียกคืน เท่ากับ 74%, ค่าเอฟวัน เท่ากับ 76%, ค่าความแม่น เท่ากับ 78% และค่าเอยูซี เท่ากับ 0.94 นอกจากนี้ค่าคะแนนประสบการณ์ผู้ใช้โดยรวมของโมไบล์แอปพลิเคชันที่คำนวณได้มีค่าสหสัมพันธ์เชิงบวกในระดับปานกลางถึงสูงเป็น 0.68 กับค่าคะแนนการจัดอันดับบนแอปสโตร์ ผลการศึกษานี้ชี้ให้เห็นถึงศักยภาพของการวิเคราะห์และประเมินคะแนนประสบการณ์ผู้ใช้อย่างอัตโนมัติในการเสริมสร้างคุณภาพของโมไบล์แอปพลิเคชันและความพึงพอใจของผู้ใช้

การพัฒนาส่วนแบ็กเอนด์ของเว็บแอปพลิเคชันโดยใช้เฟรมเวิร์กโหนดเจเอสแบบอิงแบบจำลอง, วีรยุทธ กันภัย

Chulalongkorn University Theses and Dissertations (Chula ETD)

การพัฒนาซอฟต์แวร์แบบอิงแบบจำลอง ได้กลายเป็นแนวทางที่มีศักยภาพในการวิศวกรรมซอฟต์แวร์ ซึ่งใช้แบบจำลองเป็นองค์ประกอบหลักในการสนับสนุนการพัฒนาซอฟต์แวร์ รวมถึงการพัฒนาเว็บแอปพลิเคชัน งานวิจัยนี้นำเสนอการประยุกต์ใช้แนวคิดของการพัฒนาแบบอิงแบบจำลองกับการพัฒนาแบ็กเอนด์ด้วยโหนดเจเอสโดยเน้นทั้งกระบวนการวิศวกรรมไปข้างหน้า และวิศวกรรมย้อนกลับ โดยได้พัฒนายูเอ็มแอลโพรไฟล์สำหรับโหนดเจเอสที่รองรับทั้งฐานข้อมูลแบบซีเควลและโนซีเควลเพื่อใช้ในการสร้างแบบจำลองการออกแบบของเว็บแอปพลิเคชัน และมีเครื่องมือสำหรับการแปลงแบบจำลองเป็นโค้ดและโค้ดเป็นแบบจำลอง ผลการประเมินแนวทางที่นำเสนอผ่านกรณีศึกษา 3 กรณี พบว่ากระบวนการวิศวกรรมไปข้างหน้ามีผลของอัตราการแปลงแบบจำลองเป็นโค้ดที่สูง โดยเฉพาะในส่วนของ Model (86.18-94.53%) และ Route (70.73-90.52%) โดยมีประสิทธิภาพที่ดีสอดคล้องกันทั้งในกรณีของซีเควล (มายซีเควล) และ โนซีเควล (มองโกดีบี) ขณะที่กระบวนการวิศวกรรมย้อนกลับให้ผลลัพธ์ความถูกต้องของการแปลงโค้ดเป็นแบบจำลองถึง 100% สำหรับโค้ดแบ็กเอนด์ในแต่ละรูปแบบการพัฒนา การดำเนินการนี้ช่วยลดระยะเวลาในการพัฒนาโค้ดอย่างมีนัยสำคัญ พร้อมทั้งคงไว้ซึ่งเอกสารประกอบระบบผ่านการสร้างแบบจำลองโดยอัตโนมัติ ซึ่งเป็นแนวทางที่ใช้งานได้จริงสำหรับการพัฒนาและบำรุงรักษาแบ็กเอนด์ด้วยโหนดเจเอส

การค้นหาค่าสัมประสิทธ์การปล่อยก๊าซเรือนกระจกบนพื้นฐานแบบจำลองภาษาขนาดใหญ่, สุชานันท์ กระบวนยุทธ

Chulalongkorn University Theses and Dissertations (Chula ETD)

งานวิจัยนี้ดำเนินการเพื่อวิจัยแบบจำลองภาษาขนาดใหญ่และเทคนิคต่างๆ ที่ช่วยเพิ่มศักยภาพของโมเดลเหล่านี้ รวมถึงวิศวกรรมที่รวดเร็วและการขยายความรู้เฉพาะ การศึกษานี้เกี่ยวข้องกับการออกแบบและพัฒนา AI agent เพื่อส่งคืนค่าสัมประสิทธ์ปล่อยก๊าซเรือนกระจกโดยอ้างอิงจากฐานข้อมูลที่ดูแลโดยองค์การบริหารก๊าซเรือนกระจก (องค์การมหาชน) ของประเทศไทย งานวิจัยนี้จะใช้ RAG เพื่อค้นหาข้อมูลที่ถูกจัดเก็บไว้ซึ่งมีความคล้ายคลึงกันมากที่สุดกับข้อมูลรับเข้า จากนั้น AI Agent จะส่งคืนค่าคำตอบเพียงคำตอบเดียวเดียวของค่าสัมประสิทธิ์การปล่อยก๊าซเรือนกระจกที่เกี่ยวข้องกับกิจกรรมที่รับเข้ามา จากการประเมินประสิทธิภาพของ LLM ขนาดกลางและขนาดใหญ่ที่ถูกเลือกทั้งห้าแบบจำลองนั้นมีแนวโน้มที่ดี

Causal Event Graph-Guided Language-Based Spatiotemporal Question Answering, Kaushik Roy, Alessandro Oltramari, Yuxin Zi, Chathurangi Shyalika, Vignesh Narayanan, Amit Sheth

Publications

Large Language Models have excelled at encoding and leveraging language patterns in large text-based corpora for various tasks, including spatiotemporal event-based question answering (QA). However, due to encoding a text-based projection of the world, they have also been shown to lack a fullbodied understanding of such events, e.g., a sense of intuitive physics, and cause-and-effect relationships among events. In this work, we propose using causal event graphs (CEGs) to enhance language understanding of spatiotemporal events in language models, using a novel approach that also provides proofs for the model’s capture of the CEGs. A CEG consists of events denoted by …

Advancing A Systems Perspective On Innovative Behavior, Stephen Demski

Doctoral Dissertations

"Engineering organizations pursue innovation in strategy, structure, processes, and the services and products offered to remain relevant and competitive. Identifying factors supporting or constraining innovative work behavior and recognizing the complexity of their interactions are vital to sustaining an innovative workforce, yet how factors interact has not been comprehensively studied.

Recognizing innovative work behavior as the output of a complex system of factors guided this study’s literature search that identified over one hundred individual, team, and organizational factors influencing innovative behavior, interviews of engineers to learn what factors are essential in their work environment, and Delphi survey to rank factors, …

Communication Modality And Activity Recognition In Smart Manufacturing, Haodong Chen

Doctoral Dissertations

"Advancements in sensors, computational intelligence, and big data have been rapidly transforming and revolutionizing the manufacturing sector, leading to robot-rich and digitally connected factories. To facilitate such a transformative process, workforce training and coordination between the human workforce and robots have remained a major challenge. This study aims to create intelligent workforce training systems for human-computer interaction and human-robot collaboration. It proposes the use of dynamic gestures and speech commands for seamless communication between a worker and a robot, leveraging Convolutional Neural Networks (CNN) and multiple threading for recognition and integration. Additionally, a fine-grained activity recognition model has been crafted …

Accurate And Time Efficient Signal Integrity And Power Integrity Modeling Of High-Speed Digital Systems, Chaofeng Li

Doctoral Dissertations

"Signal integrity (SI) and power integrity (PI) play an important role in the modern high-speed digital system design, which are closely related to the printed circuit board (PCB) dielectric material property, the PCB interconnect performance, and the power delivery network (PDN) on PCB. Generally, the full-wave simulation is used to accurately analyze and evaluate the designed PCB. But full-wave simulation is not a good option for the complex PCB structure with high aspect ratio, for example, PCB vias, and PDN, which will require significant computing time and storage resources. Equivalent circuit models have been developed to efficiently predict the electrical …

Effect Of Ai–Si Coating Weights On Corrosion And Weldability Of Ultra-High-Strength Steel Usiborr 1500 Used In Automotive Structures, Hilary Anayochukwu Onyishi

Doctoral Dissertations

"There has been increased demand for hot-stamped pre-coated Usibor 1500R in automotive industries due to its excellent properties (ultra-high strength, weight reduction, low spring back and excellent corrosion protection). In this research work, investigation was conducted to better understand the effect of Ai-Si Coating Weights on weldability and corrosion of Ultra-High-Strength Steel Usibor 1500R used in Automotive Structures”. The results show that weld nugget diameter increases with increasing weld current, better spot-weld nugget diameter is recorded at weld currents above 7.5 kA for both AS80 and AS150 and the weld current range decreases as the interdiffusion layer (IDL) thickness increases. …

Lightweight Neural Network Pipeline Model For Drone Collision Prediction System Using Limited Computing Resources, Rifqi Nabila Zufar

Chulalongkorn University Theses and Dissertations (Chula ETD)

This thesis focuses on developing a drone collision prediction algorithm by adapting the Neural Network Pipeline (NNP) for use in limited computing resources devices. While NNP is effective with sufficient computation, its performance in constrained environments remains untested. To address this issue, an adaptation of NNP using post-training quantization is developed. This method effectively compresses the model by reducing the precision of the 32-bit floating point. This adaptation aims at reducing model complexity and offers significant benefits in terms of reduced model size, increased processing speed, and reduced power consumption. It results in the expense of prediction accuracy. The MobileNetV3Small …