Physical Sciences and Mathematics Commons^™

Optimization And Application Of Graph Neural Networks, Shuo Zhang

Dissertations, Theses, and Capstone Projects

Graph Neural Networks (GNNs) are widely recognized for their potential in learning from graph-structured data and solving complex problems. However, optimal performance and applicability of GNNs have been an open-ended challenge. This dissertation presents a series of substantial advances addressing this problem. First, we investigate attention-based GNNs, revealing a critical shortcoming: their ignorance of cardinality information that impacts their discriminative power. To rectify this, we propose Cardinality Preserved Attention (CPA) models that can be applied to any attention-based GNNs, which exhibit a marked improvement in performance. Next, we introduce the Directional Node Pair (DNP) descriptor and the Robust Molecular Graph …

Carbon-Aware Mine Planning With A Novel Multi-Objective Framework, Nurul Asyikeen Binte Azhar, Aldy Gunawan, Shih-Fen Cheng, Erwin Leonardi

Research Collection School Of Computing and Information Systems

The logistical complication of long-term mine planning involves deciding the sequential extraction of materials from the mine pit and their subsequent processing steps based on geological, geometrical, and resource constraints. The net present value (NPV) of profit over the mine's lifespan usually forms the sole objective for this problem, which is considered as the NP-hard precedence-constrained production scheduling problem (PCPSP) as well. However, increased pressure for more sustainable and carbon-aware industries also calls for environmental indicators to be considered. In this paper, we enhance the generic PCPSP formulation into a multi-objective optimization (MOO) problem whereby carbon cost forms an additional …

Rosas: Deep Semi-Supervised Anomaly Detection With Contamination-Resilient Continuous Supervision, Hongzuo Xu, Yijie Wang, Guansong Pang, Songlei Jian, Ning Liu, Yongjun Wang

Research Collection School Of Computing and Information Systems

Semi-supervised anomaly detection methods leverage a few anomaly examples to yield drastically improved performance compared to unsupervised models. However, they still suffer from two limitations: 1) unlabeled anomalies (i.e., anomaly contamination) may mislead the learning process when all the unlabeled data are employed as inliers for model training; 2) only discrete supervision information (such as binary or ordinal data labels) is exploited, which leads to suboptimal learning of anomaly scores that essentially take on a continuous distribution. Therefore, this paper proposes a novel semi-supervised anomaly detection method, which devises contamination-resilient continuous supervisory signals. Specifically, we propose a mass interpolation method …

Constrained Multiagent Reinforcement Learning For Large Agent Population, Jiajing Ling, Arambam James Singh, Duc Thien Nguyen, Akshat Kumar

Research Collection School Of Computing and Information Systems

Learning control policies for a large number of agents in a decentralized setting is challenging due to partial observability, uncertainty in the environment, and scalability challenges. While several scalable multiagent RL (MARL) methods have been proposed, relatively few approaches exist for large scale constrained MARL settings. To address this, we first formulate the constrained MARL problem in a collective multiagent setting where interactions among agents are governed by the aggregate count and types of agents, and do not depend on agents’ specific identities. Second, we show that standard Lagrangian relaxation methods, which are popular for single agent RL, do not …

Perceptions And Barriers To Adopting Artificial Intelligence In K-12 Education: A Survey Of Educators In Fifty States, Karen Woodruff, James Hutson, Kathryn Arnone

Faculty Scholarship

Artificial Intelligence (AI) is making significant strides in the field of education, offering new opportunities for personalized learning and access to education for a more diverse population. Despite this potential, the adoption of AI in K-12 education is limited, and educators’ express hesitancy towards its integration due to perceived technological barriers and misconceptions. The purpose of this study is to examine the perceptions of K-12 educators in all 50 states of the USA towards AI, policies, training, and resources related to technology and AI, their comfort with technology, willingness to adopt new technologies for classroom instruction, and needs assessment for …

Advances In Quaternion-Valued Neural Networks, Jeremiah P. Bill

Theses and Dissertations

This dissertation investigates the construction, optimization, and application of quaternion neural networks (QNNs) to Department of Defense (DoD) related problem sets. QNNs are a type of neural network wherein the weights, biases, and input values are all represented as quaternion numbers. This work provides a critical evaluation of the myriad different quaternion backpropagation derivations that exist in the literature, testing the performance of each on a range of regression problem sets. The optimization dynamics of QNNs are explored, presenting visualizations of QNN loss surfaces and a novel method for assessing the “smoothness” of these loss surfaces. Finally, this dissertation presents …

Out-Of-Distribution Generalization Of Deep Learning To Illuminate Dark Protein Functional Space, Tian Cai

Dissertations, Theses, and Capstone Projects

Dark protein illumination is a fundamental challenge in drug discovery where majority human proteins are understudied, i.e. with only known protein sequence but no known small molecule binder. It's a major road block to enable drug discovery paradigm shift from single-targeted which looks to identify a single target and design drug to regulate the single target to multi-targeted in a Systems Pharmacology perspective. Diseases such as Alzheimer's and Opioid-Use-Disorder plaguing millions of patients call for effective multi-targeted approach involving dark proteins. Using limited protein data to predict dark protein property requires deep learning systems with OOD generalization capacity. Out-of-Distribution (OOD) …

Uncertainty-Adjusted Inductive Matrix Completion With Graph Neural Networks, Petr Kasalicky, Antoine Ledent, Rodrigo Alves

Research Collection School Of Computing and Information Systems

We propose a robust recommender systems model which performs matrix completion and a ratings-wise uncertainty estimation jointly. Whilst the prediction module is purely based on an implicit low-rank assumption imposed via nuclear norm regularization, our loss function is augmented by an uncertainty estimation module which learns an anomaly score for each individual rating via a Graph Neural Network: data points deemed more anomalous by the GNN are downregulated in the loss function used to train the low-rank module. The whole model is trained in an end-to-end fashion, allowing the anomaly detection module to tap on the supervised information available in …

Are We Ready To Embrace Generative Ai For Software Q&A?, Bowen Xu, Thanh-Dat Nguyen, Thanh Le Cong, Thong Hoang, Jiakun Liu, Kisub Kim, Chen Gong, Changan Niu, Chenyu Wang, Xuan-Bach Dinh Le, David Lo

Research Collection School Of Computing and Information Systems

Stack Overflow, the world's largest software Q&A (SQA) website, is facing a significant traffic drop due to the emergence of generative AI techniques. ChatGPT is banned by Stack Overflow after only 6 days from its release. The main reason provided by the official Stack Overflow is that the answers generated by ChatGPT are of low quality. To verify this, we conduct a comparative evaluation of human-written and ChatGPT-generated answers. Our methodology employs both automatic comparison and a manual study. Our results suggest that human-written and ChatGPT-generated answers are semantically similar, however, human-written answers outperform ChatGPT-generated ones consistently across multiple aspects, …

Quantifying Balance: Computational And Learning Frameworks For The Characterization Of Balance In Bipedal Systems, Kubra Akbas

Dissertations

In clinical practice and general healthcare settings, the lack of reliable and objective balance and stability assessment metrics hinders the tracking of patient performance progression during rehabilitation; the assessment of bipedal balance plays a crucial role in understanding stability and falls in humans and other bipeds, while providing clinicians important information regarding rehabilitation outcomes. Bipedal balance has often been examined through kinematic or kinetic quantities, such as the Zero Moment Point and Center of Pressure; however, analyzing balance specifically through the body's Center of Mass (COM) state offers a holistic and easily comprehensible view of balance and stability.

Building upon …

Learning Representations For Effective And Explainable Software Bug Detection And Fixing, Yi Li

Dissertations

Software has an integral role in modern life; hence software bugs, which undermine software quality and reliability, have substantial societal and economic implications. The advent of machine learning and deep learning in software engineering has led to major advances in bug detection and fixing approaches, yet they fall short of desired precision and recall. This shortfall arises from the absence of a 'bridge,' known as learning code representations, that can transform information from source code into a suitable representation for effective processing via machine and deep learning.

This dissertation builds such a bridge. Specifically, it presents solutions for effectively learning …

Fortifying Robustness: Unveiling The Intricacies Of Training And Inference Vulnerabilities In Centralized And Federated Neural Networks, Guanxiong Liu

Dissertations

Neural network (NN) classifiers have gained significant traction in diverse domains such as natural language processing, computer vision, and cybersecurity, owing to their remarkable ability to approximate complex latent distributions from data. Nevertheless, the conventional assumption of an attack-free operating environment has been challenged by the emergence of adversarial examples. These perturbed samples, which are typically imperceptible to human observers, can lead to misclassifications by the NN classifiers. Moreover, recent studies have uncovered the ability of poisoned training data to generate Trojan backdoored classifiers that exhibit misclassification behavior triggered by predefined patterns.

In recent years, significant research efforts have been …

Ai-Supported Academic Advising: Exploring Chatgpt’S Current State And Future Potential Toward Student Empowerment, Daisuke Akiba, Michelle C. Fraboni

Publications and Research

Artificial intelligence (AI), once a phenomenon primarily in the world of science fiction, has evolved rapidly in recent years, steadily infiltrating into our daily lives. ChatGPT, a freely accessible AI-powered large language model designed to generate human-like text responses to users, has been utilized in several areas, such as the healthcare industry, to facilitate interactive dissemination of information and decision-making. Academic advising has been essential in promoting success among university students, particularly those from disadvantaged backgrounds. Unfortunately, however, student advising has been marred with problems, with the availability and accessibility of adequate advising being among the hurdles. The current study …

Military Metaverse: Key Technologies, Potential Applications And Future Directions, Zhao Zhang, Yujie Guo, Xiaoning Zhao, Baoliang Sun

Journal of System Simulation

Abstract: Since its emergence, the concept of metaverse has been applied to many fields. At present, the transformation of military intelligence, digitalization and information technology is advancing in an allround way. The military enabled by metaverse technology will accelerate the process of military reform in the new era. To explore the potential application of metaverse in the military field, this paper first introduces several key technologies of metaverse and their functions in the military field, and then discusses some potential directions for the application of metaverse in weapon development and support, training and teaching of officers and soldiers, tactical command …

The Library & Generative Ai, Nat Gustafson-Sundell, Mark Mccullough

Library Services Publications

A demonstration of several AI tools, including ChatGPT, ChatPDF, Consensus, and more. The focus of the session is on potential student uses of the tools and related library initiatives, so we address the limits of ChatGPT as an information source. Librarians can help students learn how to use these tools responsibly and provide leadership on campus as AI is integrated into assignments.

Short-Term Vehicle Speed Prediction With Spatiotemporal Convolution Fused With Variational Modal Decomposition, Kai Zhang, Haipeng Lu, Ying Han, Lingyun Zhang, Yujie Ding

Journal of System Simulation

Abstract: Accurate short-term vehicle speed prediction helps to resolve city traffic congestion problems. Focusing on the defect that CNN cannot process non-Euclidean geometric data, GCN and BiLSTM are combined to fully process the spatiotemporal characteristics of road network information, in which the advantages of GCN integrating global features and the ability of BiLSTM to extract temporal features are considered. In order to reduce the interference of noise to the data, variational modal decomposition (VMD) is introduced and short-term vehicle speed prediction model based on VMD-GCN-BiLSTM (VGBLSTM) is proposed . Simulation results show that the prediction accuracy of VGBLSTM model is …

Intelligent Path Planning For Mobile Robots Based On Sac Algorithm, Laiyi Yang, Jing Bi, Haitao Yuan

Journal of System Simulation

Abstract: Aiming at the high dimension, slow convergence and complex modelling of traditional path planning algorithms for mobile robots, a new intelligent path planning algorithm is proposed, which is based on deep reinforcement learning soft actor-critic (SAC) algorithm to save the poor performance of robot in complicated environments with static and dynamic obstacles. An improved reward function is designed to enable mobile robots to quickly avoid obstacles and reach targets by using state dynamic normalization and priority experience pool techniques. To evaluate the performance, a pygame-based simulation environment is constructed. Compared with proximal policy optimization(PPO) algorithm, experimental …

A Framework On Equipment Digital Twin Credibility Assessment, Han Lu, Lin Zhang, Kunyu Wang, Zejun Huang

Journal of System Simulation

Abstract: A key bottleneck in the large-scale application of equipment digital twin is the lack of systematic and effective credibility assessment methods. This paper analyzes the dynamic evolution, virtual-real interactivity and other key features of the equipment digital twin. A credibility assessment framework for the equipment digital twin is proposed, including the credibility connotation of the digital twin, a multi-dimensional and multi-level credibility assessment index system and a credibility assessment methodology. The whole assessment process is illustrated with the robotic arm digital twin as an example, which can provide directional guidance for the assessment and construction of the digital twin.

Detection Of False Data Injection Attack In Smart Grid Based On Improved Ukf, Lisheng Wei, Qian Zhang

Journal of System Simulation

Abstract: Due to the disruption and threat of false data injection attack (FDIA) on grid cyber-physical systems (GCPS), and to address the problem that false data is difficult to be detected, a method for smart grid false data detection based on weighted least squares (WLS) and improved unscented Kalman filter (UKF) is proposed. FDIA is modeled mathematically, and the residual analysis shows that the FDIA is difficult to be detected. In the case of the injection attack vector, the improved UKF is applied to state estimation. Meanwhile, the state estimation of the system is performed by the WLS, which is …

Modeling And System Realization Of Assembly Robot Based On Digital Twin, Jian Xu, Xin Song, Xiuping Liu, Bo Chen

Journal of System Simulation

Abstract: Aiming at the problems of time-consuming, labor-intensive and poor accuracy in programming of complex operations of industrial assembly robots, an online programming method for robots based on digital twin is proposed. From the four dimensions of geometry, contact dynamics, behavior and rules, the faithful mapping of the robot from real to virtual space is realized. The verification of multi-part assembly trajectory planning program, real-time synchronous operation and state monitoring are realized. An assembly robot modeling and online programming system based on digital twin is designed and built. A six-axis industrial robot cell is taken as an example to verify …

Target Search Planning And Algorithm For Monitoring Of Polar Disaster Areas, Fei Ding, Meinan Zhang, Hengheng Zhuang, Hairong Ma

Journal of System Simulation

Abstract: Aiming at improving the ability of safe navigation route planning and risk assessment of ships in polar waters, a target search model and method based on clustering and efficient indexing of monitoring center are proposed. By constructing a disaster monitoring scenario based on the current navigation area of the ship, a virtual electronic fence is introduced to define the monitoring area. Spectral clustering algorithm is used to divide the risk level of the fence area, extract high-risk areas, and optimize the generation of target search scenarios; Efficient determination of the matching relationship between the target vessel and the fence …

Attack Decision-Making Model Of Armed Helicopter Based On Multi-Index Fuzzy Set, Chunyan Wang, Xiang Wang, Minchi Kuang, Danfeng Wu, Zhengtong Li

Journal of System Simulation

Abstract: Aiming at the attack decision-making task requirements of armed helicopters in uncertain battlefield environment, the constructed multi-index fuzzy set is quantitatively characterized by the improved Gaussian model. The strategic benefit value model is constructed by using the combat restraint relationship, and the target ranking set is obtained by dynamically assigning the weight factors of the threat value and the strategic benefit value to complete the attack decision-making. The results show that the proposed method can better use the threat index data in modeling, and can provide theoretical guidance and modeling reference to improve the decision-making advantage of armed …

Real-Time Simulation Method Of Ultra-High-Definition Video Texture, Yangyang Liu, Gangyi Ding, Dapeng Yan, Tong Xue

Journal of System Simulation

Abstract: With the development and promotion of ultra-high-definition video technology, how to quickly simulate ultra-high-definition video texture has gradually become an important research issue. Aiming at the completeness and high efficiency of simulation, a real-time simulation method of ultra-high-definition video texture is proposed to improve the video texture quality and display frequency simultaneously. A fast generation method of video texture based on GPU parallel is designed, which solves the time-consuming problem of decoding and transcoding. An efficient data transmission method based on shared texture is proposed. On the basis of the simulation engine, the real-time simulation system of ultra-high-definition video …

Modeling And Identification Of Wind Power Generation System Based On Hammerstein Model, Feng Li, Tian Zheng, Wei Song

Journal of System Simulation

Abstract: A modeling and identification method of wind power generation system based on Hammerstein model is studied to establish high-precision model of wind power generation system. Firstly, 3σ criterion is used to propose the abnormal data, and the eliminated data is used to train the nominal model of the wind power generation system. Furthermore, the Hammerstein model is used to establish the data-driven model of wind power generation system, and the combined signal composed of separable signal and actual wind speed is used as the input of the Hammerstein model. The output of the separable signal through the nominal model …

Research And Application Progress Of Tracking Registration Methods In Ar Assembly, Wei Fang, Shuhong Xu, Lei Han, Zhangwenchi Li

Journal of System Simulation

Abstract: Augmented reality (AR) can superimpose virtual auxiliary information in appropriate positions at the actual work site to achieve intelligent assembly guidance of "what you see is what you operate", alleviating the difficulties for workers to recognize drawings in traditional drawing-based assembly and the problems of misassembly, missing parts and so on. Stable tracking registration in the assembly site environment is the basis for achieving the fusion of virtual and actual scene in augmented assembly visual guidance, and also a key issue in the practical application and deployment of existing augmented assembly. In view of the research and application results …

Research On Dynamic Simulation Technology For Satellite Internet, Xiaofeng Wang, Taiqian Shen, Yuan Liu

Journal of System Simulation

Abstract: Network simulation is an important support for new technology verification and security technology evaluation in the rolling construction of satellite internet. A dynamic simulation architecture which based on cloud platform is proposed to address the highly time-varying characteristics of satellite internet. An algorithm of distributed simulation technology of satellite links, which improves the synchronization of accuracy of dynamic simulation of links, is designed to tackle the time-varying of satellite links. Aiming at the topology changes in the process of satellites motion, an algorithm of realtime topology simulation based on time splice is raised, which realizes real-time and accurate changes …

Energy Management Strategy Of Multi-Agent Microgrid Based On Q-Learning Algorithm, Miaomiao Ma, Lipeng Dong, Xiangjie Liu

Journal of System Simulation

Abstract: This paper proposes a multi-agent microgrid energy management method for the energy trading and benefit distribution in the microgrid power market based on the Q-learning algorithm. Based on the electricity market, microgrid system and transaction process are constructed to clarify the responsibilities of each unit. The mathematical models of distributed power generations are established by considering the changes in wind speed, light intensity and ambient temperature, as well as the upper and lower limit constraints of the output power of each power generation unit. On this basis, the distributed power generations and user loads are regarded as agents, and …

An Intelligent Driver Model Simulation Considering Both Backward Looking Effect And Velocity Difference, Yin Xu, Yun Pu, Haixu Liu, Yifan Tan

Journal of System Simulation

Abstract: Aiming at the phenomenon that driver adjusts vehicle movement by observing the following vehicles through rearview mirror in the actual car-following driving, an improved intelligent driver model accounting for both backward looking effect and velocity difference is proposed, and the critical stability condition of the new model is obtained by employing the linear stability analysis. Based on the numerical simulation experiments, the car following characteristics analysis during the acceleration process of the vehicle and the traffic safety evaluation are carried out. A small disturbance simulation under the periodic boundary condition is used to verify the conclusion consistency of stability …

Surface Defect Detection Of Power Equipment Using Adaptive Receptive Field Network, Hao Yu, Jinxia Jiang, Xiaohan Lai, Feng Mei

Journal of System Simulation

Abstract: For the detection of defects such as icing, rust, and contamination of power equipment in substations, a novel adaptive receptive field network (ARFN) is proposed, in which an adaptive receptive field module (ARFM) combined with the attention mechanism can effectively fuse multi-scale features. Considering the small sample learning attribute of defect detection, a power equipment surface defect simulation data synthesis method based on real texture is also proposed. The experimental results on the simulation dataset show that the network has high detection accuracy for surface defects across devices, while having advantages such as small size and fast operation speed.

Two New Maneuvering Target Simulation Methods, Yingxuan Li, Zhongxun Wang, Yunlong Dong

Journal of System Simulation

Abstract: To verify the performance of maneuvering target tracking algorithm, it's necessary to build a complex motion simulation model similar to the actual target motion situation. A simulation model of maneuvering target with controllable time correlation coefficient is constructed based on the idea of Singer model, and the suitability of Singer's maneuvering target tracking algorithm is verified when the time correlation coefficient does not match. Aiming at the problem that the traditional coordinated turning model only considers the change of normal acceleration, and the tangential acceleration is always assumed to be 0, which is not highly consistent with the actual …