Computer Engineering Commons^™

Targeted Adversarial Attacks Against Neural Network Trajectory Predictors, Kaiyuan Tan

McKelvey School of Engineering Graduate Student Theses & Dissertations

Trajectory prediction is an integral component of modern autonomous systems as it allows for envisioning future intentions of nearby moving agents. Due to the lack of other agents' dynamics and control policies, deep neural network (DNN) models are often employed for trajectory forecasting tasks. Although there exists an extensive literature on improving the accuracy of these models, there is a very limited number of works studying their robustness against adversarially crafted input trajectories. To bridge this gap, in this paper, we propose a targeted adversarial attack against DNN models for trajectory forecasting tasks. We call the proposed attack TA4TP for …

Enabling The Integration Of Sustainable Design Methodological Frameworks And Computational Life Cycle Assessment Tools Into Product Development Practice, Tejaswini Chatty

Dartmouth College Ph.D Dissertations

Environmental sustainability has gained critical importance in product development (PD) due to increased regulation, market competition, and consumer awareness, leading companies to set ambitious climate targets . To meet these goals, PD practitioners (engineers and designers) are often left to adapt their practices to reduce the impacts of the products they manufacture. Literature review and interviews with practitioners show that they highly valued using quantitative life cycle assessment (LCA) results to inform decision making.

LCA is a technique to measure the environmental impacts across various stages of a product life cycle. Existing LCA software tools, however, are designed for dedicated …

Adversarial Patch Attacks On Deep Reinforcement Learning Algorithms, Peizhen Tong

McKelvey School of Engineering Graduate Student Theses & Dissertations

Adversarial patch attack has demonstrated that it can cause the misclassification of deep neural networks to the target label when the size of patch is relatively small to the size of input image; however, the effectiveness of adversarial patch attack has never been experimented on deep reinforcement learning algorithms. We design algorithms to generate adversarial patches to attack two types of deep reinforcement learning algorithms, including deep Q-networks (DQN) and proximal policy optimization (PPO). Our algorithms of generating adversarial patch consist of two parts: choosing attack position and training adversarial patch on that position. Under the same bound of total …

Reinforcement Learning-Based Resilience And Decision Making In Cyber-Physical Systems, Fisayo Sangoleye

Electrical and Computer Engineering ETDs

Cyber-physical systems (CPS) transform how humans interact with technology by integrating sensing, computation, networking, and control with physical processes to facilitate smart services and innovative applications in our environments. Recent advances in CPS have led to rapid growth in the amount of information constantly generated by people, systems, and processes. Most of this information, however, is underutilized due to the lack of efficient information utilization and decision-making techniques. Also, the increasing interconnectivity of CPSs presents security risks that, if left unaddressed, could be highly disruptive to systems, processes, and economies. In this dissertation, we present a study and proposal of …

Software-Defined Networking Security Techniques And The Digital Forensics Of The Sdn Control Plane, Abdullah Alshaya

LSU Doctoral Dissertations

Software-Defined Networking (SDN) is an efficient networking design that decouples the network's control plane from the data plane. When compared to the traditional network architecture, the SDN architecture shares many of the same security issues. The centralized SDN controller makes it easier to control, easier to program in real-time, and more flexible, but this comes at the cost of more security risks. An attack on the control plane layer of the SDN controller is a major security concern.

First, centralized design and the existence of a single point of failure in the control plane compromise the accessibility and availability of …

Grammatical Triples Extraction For The Distant Reading Of Textual Corpora, Stephanie Buongiorno, Stephanie Buongiorno

Multidisciplinary Studies Theses and Dissertations

Grammatical triples extraction has become increasingly important for the analysis of large, textual corpora. By providing insight into the sentence-level linguistic features of a corpus, extracted triples have supported interpretations of some of the most relevant problems of our time. The growing importance of triples extraction for analyzing large corpora has put the quality of extracted triples under new scrutiny, however. Triples outputs are known to have large amounts of erroneous triples. The extraction of erroneous triples poses a risk for understanding a textual corpus because erroneous triples can be nonfactual and even analogous to misinformation. Disciplines such as the …

Vi Energy-Efficient Memristor-Based Neuromorphic Computing Circuits And Systems For Radiation Detection Applications, Jorge Iván Canales Verdial

Electrical and Computer Engineering ETDs

Radionuclide spectroscopic sensor data is analyzed with minimal power consumption through the use of neuromorphic computing architectures. Memristor crossbars are harnessed as the computational substrate in this non-conventional computing platform and integrated with CMOS-based neurons to mimic the computational dynamics observed in the mammalian brain’s visual cortex. Functional prototypes using spiking sparse locally competitive approximations are presented. The architectures are evaluated for classification accuracy and energy efficiency. The proposed systems achieve a 90% true positive accuracy with a high-resolution detector and 86% with a low-resolution detector.

Secure And Efficient Federated Learning, Xingyu Li

Theses and Dissertations

In the past 10 years, the growth of machine learning technology has been significant, largely due to the availability of large datasets for training. However, gathering a sufficient amount of data on a central server can be challenging. Additionally, with the rise of mobile networking and the large amounts of data generated by IoT devices, privacy and security issues have become a concern, resulting in government regulations such as GDPR, HIPAA, CCPA, and ADPPA. Under these circumstances, traditional centralized machine learning methods face a problem in that sensitive data must be kept locally for privacy reasons, making it difficult to …

Unmanned Aerial System Integration Safety And Security Technology Ontology, Rebecca A. Garcia

Theses and Dissertations

Unmanned Aerial System (UAS) is a versatile and essential tool for law enforcement, first responders, utility providers, and the general public. Integrating the UAS into the National Airspace System (NAS) poses a significant challenge to policymakers and manufacturers. A UAS Integration Safety and Security Technology Ontology (ISSTO) has been developed in the Web Ontology Language (OWL) to aid in this integration. ISSTO is a domain ontology covering aviation topics corresponding to flights, aircraft types, manufacturers, temporal/spatial, waivers and authorizations, track data, NAS facilities, air traffic control advisories, weather phenomena, surveillance and security equipment, and events, sensor types, radio frequency ranges, …

Enabling Security Analysis And Education Of The Ethereum Platform: A Network Traffic Dissection Tool, Joshua Mason Kemp

Masters Theses, 2020-current

Ethereum, the decentralized global software platform powered by blockchain technology known for its native cryptocurrency, Ether (ETH), provides a technology stack for building apps, holding assets, transacting, and communicating without control by a central authority. At the core of Ethereum’s network is a suite of purpose-built protocols known as DEVP2P, which provides the underlying nodes in an Ethereum network the ability to discover, authenticate and communicate confidentiality. This document discusses the creation of a new Wireshark dissector for DEVP2P’s discovery protocols, DiscoveryV4 and DiscoveryV5, and a dissector for RLPx, an extensible TCP transport protocol for a range of Ethereum node …

Sensitive And Makeable Computational Materials For The Creation Of Smart Everyday Objects, Te-Yen Wu Mr

Dartmouth College Ph.D Dissertations

The vision of computational materials is to create smart everyday objects using the materi- als that have sensing and computational capabilities embedded into them. However, today’s development of computational materials is limited because its interfaces (i.e. sensors) are unable to support wide ranges of human interactions , and withstand the fabrication meth- ods of everyday objects (e.g. cutting and assembling). These barriers hinder citizens from creating smart every day objects using computational materials on a large scale.

To overcome the barriers, this dissertation presents the approaches to develop compu- tational materials to be 1) sensitive to a wide variety of …

Detection Of Crypto-Ransomware Attack Using Deep Learning, Muna Jemal

Master of Science in Computer Science Theses

The number one threat to the digital world is the exponential increase in ransomware attacks. Ransomware is malware that prevents victims from accessing their resources by locking or encrypting the data until a ransom is paid. With individuals and businesses growing dependencies on technology and the Internet, researchers in the cyber security field are looking for different measures to prevent malicious attackers from having a successful campaign. A new ransomware variant is being introduced daily, thus behavior-based analysis of detecting ransomware attacks is more effective than the traditional static analysis. This paper proposes a multi-variant classification to detect ransomware I/O …

Efficient Approaches For Qubit Mapping On Nisq Computers, Sri Sesha Sailaja Lakshmi Tulasi Khandavilli

Master of Science in Computer Science Theses

—Quantum computing is gaining momentum in revolutionizing the way we approach complex problem-solving. However, the practical implementation of quantum algorithms remains a significant challenge due to the error-prone and hardware limits of near-term quantum devices. For instance, physical qubit connections are limited, which necessitates the use of quantum SWAP gates to dynamically transform the logical topology during execution. In addition, to optimize fidelity, it is essential to ensure that 1) the allocated hardware has a low error rate and 2) the number of SWAP gates injected into the circuit is minimized. To address these challenges, we propose a suite of …

When Ai Moves Downstream, Frances S. Grodzinsky, Keith W. Miller, Marty J. Wolf

School of Computer Science & Engineering Faculty Publications

After computing professionals design, develop, and deploy software, what is their responsibility for subsequent uses of that software “downstream” by others? Furthermore, does it matter ethically if the software in question is considered to be artificial intelligent (AI)? The authors have previously developed a model to explore downstream accountability, called the Software Responsibility Attribution System (SRAS). In this paper, we explore three recent publications relevant to downstream accountability, and focus particularly on examples of AI software. Based on our understanding of the three papers, we suggest refinements of SRAS.

Predicting Suicide Risk Among Youths Using Machine Learning Methods, Saswati Bhattacharjee

Master's Theses

Suicide is the second leading cause of death among youths in the USA. Although machine learning approaches have provided great potential for predicting suicide risk using survey data, prediction accuracy may not meet the need for clinical diagnosis due to the intrinsic characteristics of datasets. In this study, I perform a comparative study of six classification algorithms including naïve Bayes (NB), logistic regression (LR), multilayer perceptron (MLP), AdaBoost (Ada), random forest (RF), and bagging using YRBSS dataset and investigate the effectiveness of several data handling techniques to improve the overall performance of suicide risk prediction.

The dataset consists of 76 …

Blockchain Security: Double-Spending Attack And Prevention, William Henry Scott Iii

Electronic Theses and Dissertations

This thesis shows that distributed consensus systems based on proof of work are vulnerable to hashrate-based double-spending attacks due to abuse of majority rule. Through building a private fork of Litecoin and executing a double-spending attack this thesis examines the mechanics and principles behind the attack. This thesis also conducts a survey of preventative measures used to deter double-spending attacks, concluding that a decentralized peer-to-peer network using proof of work is best protected by the addition of an observer system whether internal or external.

Is Realt Reality? Investigating The Use Of Blockchain Technology And Tokenization In Real Estate Transactions, Caroline Moriarty

Minnesota Journal of Law, Science & Technology

No abstract provided.

An Investigation On The Resilience Of Long Short-Term Memory Deep Neural Networks, Christopher Vasquez

LSU Master's Theses

In a world of continuously advancing technology, the reliance on these technologies continues to increase. Recently, transformer networks [22] have been implemented through various projects such as ChatGPT. These networks are extremely computationally demanding and require cutting-edge hardware to explore. However, with the growing increase and popularity of these neural networks, a question of reliability and resilience comes about, especially as the dependency and research on these networks grow. Given the computational demand of transformer networks, we investigate the resilience of the weights and biases of the predecessor of these networks, i.e. the Long Short-Term (LSTM) neural network, through four …

Osmosis: Asymmetries In Telematic Performance, Matthias Ziegler

Journal of Network Music and Arts

For the 2022 edition of the NowNet Arts Conference on October 31st, the telematic research team of the Zurich University of the Arts (ZHdK) presented a project entitled OSMOSIS. OSMOSIS was a concert event, highlighting how telematically connected spaces always confront each other asymmetrically. Their telematic connection is part of a continuous space in which information is fragmented and selectively reassembled. Like the biochemical process of osmosis in which molecules diffuse across a cell membrane from one level of concentration to another, in telematic connections certain elements such as sound, physicality, movement, and empathy are diffused across spaces, each being …

Synthesis: Works Of Sarah Weaver And Collaborations (2020-2022), Sarah Weaver

Journal of Network Music and Arts

Synthesis Series is a set of contemplative contemporary network arts works for solo, chamber, and large ensemble performance. The works are my compositions and collaborations from the years 2020 to 2022. The concept of synthesis is conceived as an activation of synchrony. Synthesis Series follows my prior works in Synchrony Series and Source Series as sequences of compositions since 1998 proliferating into a networked system of artistic realization. In Synchrony Series I defined synchrony as the perceptual alignment of distributed time and space components. Synthesis builds on this to activate the alignment as a networked state of composite resultants, networked …