Physical Sciences and Mathematics Commons^™

The Vulnerabilities Of Artificial Intelligence Models And Potential Defenses, Felix Iov

Cybersecurity Undergraduate Research Showcase

The rapid integration of artificial intelligence (AI) into various commercial products has raised concerns about the security risks posed by adversarial attacks. These attacks manipulate input data to disrupt the functioning of AI models, potentially leading to severe consequences such as self-driving car crashes, financial losses, or data breaches. We will explore neural networks, their weaknesses, and potential defenses. We will discuss adversarial attacks including data poisoning, backdoor attacks, evasion attacks, and prompt injection. Then, we will explore defense strategies such as data protection, input sanitization, and adversarial training. By understanding how adversarial attacks work and the defenses against them, …

High-Resolution And Quality Settings With Latent Consistency Models, Steven Chen, Junrui Zhang, Rui Ning

Cybersecurity Undergraduate Research Showcase

Diffusion Models have become powerful generative models which is capable of synthesizing high-quality images across various domains. This paper explores Stable Diffusion and mostly focuses on Latent Diffusion Models. Latent Consistency Models can enhance the inference with minimal iterations. It demonstrates the performance in image in-painting and class-conditional synthesis tasks. Throughout the experiment different datasets and parameter configurations, the paper highlights the image quality, processing time, and parameter. It also discussed the future directions including adding trigger-based implementation and emotional-based themes to replace the prompt.

The Security Of Deep Neural Networks, Jalaya Allen

Cybersecurity Undergraduate Research Showcase

Our society has transitioned from our primitive lifestyle to soon, an increasingly automatic one. That idea is further exemplified as we shift into an AI era, better known as Artificial intelligence. Artificial Intelligence is classified as computer systems that can perform tasks that typically require human intelligence. However, a common thought or question that most might have is, how is this done? How does AI process information the way we want it to and have access to so much information? AI is trained by systems called AI models. These modeling programs are trained on data to recognize patterns or make …

The Transformative Integration Of Artificial Intelligence With Cmmc And Nist 800-171 For Advanced Risk Management And Compliance, Mia Lunati

Cybersecurity Undergraduate Research Showcase

This paper explores the transformative potential of integrating Artificial Intelligence (AI) with established cybersecurity frameworks such as the Cybersecurity Maturity Model Certification (CMMC) and the National Institute of Standards and Technology (NIST) Special Publication 800-171. The thesis argues that the relationship between AI and these frameworks has the capacity to transform risk management in cybersecurity, where it could serve as a critical element in threat mitigation. In addition to addressing AI’s capabilities, this paper acknowledges the risks and limitations of these systems, highlighting the need for extensive research and monitoring when relying on AI. One must understand boundaries when integrating …

The Analysis And Impact Of Artificial Intelligence On Job Loss, Ava Baratz

Cybersecurity Undergraduate Research Showcase

This paper illustrates the analysis and impact of Artificial Intelligence (AI) on job loss across various industries. This paper will discuss an overview of AI technology, a brief history of AI in industry, the positive impacts of AI, the negative impacts of AI on employment, AI considerations that contribute to job loss, the future outlook of AI, and employment loss mitigation strategies Various professional source articles and reputable blog posts will be used to finalize research on this topic.

Rising Threat - Deepfakes And National Security In The Age Of Digital Deception, Dougo Kone-Sow

Cybersecurity Undergraduate Research Showcase

This paper delves into the intricate landscape of deepfakes, exploring their genesis, capabilities, and far-reaching implications. The rise of deepfake technology presents an unprecedented threat to American national security, propagating disinformation and manipulation across various media formats. Notably, deepfakes have evolved from a historical backdrop of disinformation campaigns, merging with the advancements of artificial intelligence (AI) and machine learning to craft convincing but false multimedia content.

Examining the capabilities of deepfakes reveals their potential for misuse, evidenced by instances targeting individuals, companies, and even influencing political events like the 2020 U.S. elections. The paper highlights the direct threats posed by …

Integrating Ai Into Uavs, Huong Quach

Cybersecurity Undergraduate Research Showcase

This research project explores the application of Deep Learning (DL) techniques, specifically Convolutional Neural Networks (CNNs), to develop a smoke detection algorithm for deployment on mobile platforms, such as drones and self-driving vehicles. The project focuses on enhancing the decision-making capabilities of these platforms in emergency response situations. The methodology involves three phases: algorithm development, algorithm implementation, and testing and optimization. The developed CNN model, based on ResNet50 architecture, is trained on a dataset of fire, smoke, and neutral images obtained from the web. The algorithm is implemented on the Jetson Nano platform to provide responsive support for first responders. …

New Paths Of Attacks: Revealing The Adaptive Integration Of Artificial Intelligence In Evolving Cyber Threats Targeting Social Media Users And Their Data, Larry Teasley

Cybersecurity Undergraduate Research Showcase

The intersection between artificial intelligence tools and social media has opened doors to numerous opportunities and risks. This research delves into the escalating threat landscape in a society heavily dependent on social media. Despite the efforts by social media companies and cybersecurity professionals to mitigate cyber-attacks, the constant advancements of new technologies render social media platforms increasingly vulnerable. Malicious actors exploit generative AI to collect user data, enhancing cyber threats on social media. Notably, generative AI amplifies phishing attacks, disseminates false information, and propagates propaganda, posing substantial challenges to platform security. Ease access to large language models (LLMs) further complicates …

Knowing Just Enough To Be Dangerous: The Sociological Effects Of Censoring Public Ai, David Hopkins

Cybersecurity Undergraduate Research Showcase

This paper will present the capabilities and security concerns of public AI, also called generative AI, and look at the societal and sociological effects of implementing regulations of this technology.

What Effects Do Large Language Models Have On Cybersecurity, Josiah Marshall

Cybersecurity Undergraduate Research Showcase

Large Language Models (LLMs) are artificial intelligence (AI) tools that can process, summarize, and translate texts and predict future words in a sentence, letting the LLM generate sentences similar to how humans talk and write. One concern that needs to be flagged is that, often, the content generated by different LLMs is inaccurate. LLMs are trained on code that can be used to detect data breaches, detect ransomware, and even pinpoint organizational vulnerabilities in advance of a cyberattack. LLMs are new but have unbelievable potential with their ability to generate code that brings awareness to cyber analysts and IT professionals. …

Visual Art In The Age Of Ai, Roshnica Gurung

Cybersecurity Undergraduate Research Showcase

Artists and researchers have been deeply interested in using AI programs that generate art for quite some time now. As a result, there have been many advancements in making AI more accessible and easier to use for the public. This is because AI is not just for business anymore. Nowadays an individual without a college degree with even the slightest interest in art can go on a website like Stable Diffusion and create an artistic image using a text prompt in a quick couple minutes. The only limit is your imagination- and your internet’s stability. This accessibility was a huge …

Lidar Buoy Detection For Autonomous Marine Vessel Using Pointnet Classification, Christopher Adolphi, Dorothy Dorie Parry, Yaohang Li, Masha Sosonkina, Ahmet Saglam, Yiannis E. Papelis

Modeling, Simulation and Visualization Student Capstone Conference

Maritime autonomy, specifically the use of autonomous and semi-autonomous maritime vessels, is a key enabling technology supporting a set of diverse and critical research areas, including coastal and environmental resilience, assessment of waterway health, ecosystem/asset monitoring and maritime port security. Critical to the safe, efficient and reliable operation of an autonomous maritime vessel is its ability to perceive on-the-fly the external environment through onboard sensors. In this paper, buoy detection for LiDAR images is explored by using several tools and techniques: machine learning methods, Unity Game Engine (herein referred to as Unity) simulation, and traditional image processing. The Unity Game …

Behind Derogatory Migrants' Terms For Venezuelan Migrants: Xenophobia And Sexism Identification With Twitter Data And Nlp, Joseph Martínez, Melissa Miller-Felton, Jose Padilla, Erika Frydenlund

Modeling, Simulation and Visualization Student Capstone Conference

The sudden arrival of many migrants can present new challenges for host communities and create negative attitudes that reflect that tension. In the case of Colombia, with the influx of over 2.5 million Venezuelan migrants, such tensions arose. Our research objective is to investigate how those sentiments arise in social media. We focused on monitoring derogatory terms for Venezuelans, specifically veneco and veneca. Using a dataset of 5.7 million tweets from Colombian users between 2015 and 2021, we determined the proportion of tweets containing those terms. We observed a high prevalence of xenophobic and defamatory language correlated with the …

Statistical Approach To Quantifying Interceptability Of Interaction Scenarios For Testing Autonomous Surface Vessels, Benjamin E. Hargis, Yiannis E. Papelis

Modeling, Simulation and Visualization Student Capstone Conference

This paper presents a probabilistic approach to quantifying interceptability of an interaction scenario designed to test collision avoidance of autonomous navigation algorithms. Interceptability is one of many measures to determine the complexity or difficulty of an interaction scenario. This approach uses a combined probability model of capability and intent to create a predicted position probability map for the system under test. Then, intercept-ability is quantified by determining the overlap between the system under test probability map and the intruder’s capability model. The approach is general; however, a demonstration is provided using kinematic capability models and an odometry-based intent model.

Towards Nlp-Based Conceptual Modeling Frameworks, David Shuttleworth, Jose Padilla

Modeling, Simulation and Visualization Student Capstone Conference

This paper presents preliminary research using Natural Language Processing (NLP) to support the development of conceptual modeling frameworks. NLP-based frameworks are intended to lower the barrier of entry for non-modelers to develop models and to facilitate communication across disciplines considering simulations in research efforts. NLP drives conceptual modeling in two ways. Firstly, it attempts to automate the generation of conceptual models and simulation specifications, derived from non-modelers’ narratives, while standardizing the conceptual modeling process and outcome. Secondly, as the process is automated, it is simpler to replicate and be followed by modelers and non-modelers. This allows for using a common …

Enhancing Pedestrian-Autonomous Vehicle Safety In Low Visibility Scenarios: A Comprehensive Simulation Method, Zizheng Yan, Yang Liu, Hong Yang

Modeling, Simulation and Visualization Student Capstone Conference

Self-driving cars raise safety concerns, particularly regarding pedestrian interactions. Current research lacks a systematic understanding of these interactions in diverse scenarios. Autonomous Vehicle (AV) performance can vary due to perception accuracy, algorithm reliability, and environmental dynamics. This study examines AV-pedestrian safety issues, focusing on low visibility conditions, using a co-simulation framework combining virtual reality and an autonomous driving simulator. 40 experiments were conducted, extracting surrogate safety measures (SSMs) from AV and pedestrian trajectories. The results indicate that low visibility can impair AV performance, increasing conflict risks for pedestrians. AV algorithms may require further enhancements and validations for consistent safety performance …

Role Of Ai In Threat Detection And Zero-Day Attacks, Kelly Morgan

Cybersecurity Undergraduate Research Showcase

Cybercrime and attack methods have been steadily increasing since the 2019 pandemic. In the years following 2019, the number of victims and attacks per hour rapidly increased as businesses and organizations transitioned to digital environments for business continuity amidst lockdowns. In most scenarios cybercriminals continued to use conventional attack methods and known vulnerabilities that would cause minimal damage to an organization with a robust cyber security posture. However, zero-day exploits have skyrocketed across all industries with an increasingly growing technological landscape encompassing internet of things (IoT), cloud hosting, and more advanced mobile technologies. Reports by Mandiant Threat Intelligence (2022) concluded …

Leveraging Artificial Intelligence And Machine Learning For Enhanced Cybersecurity: A Proposal To Defeat Malware, Emmanuel Boateng

Cybersecurity Undergraduate Research Showcase

Cybersecurity is very crucial in the digital age in order to safeguard the availability, confidentiality, and integrity of data and systems. Mitigation techniques used in the industry include Multi-factor Authentication (MFA), Incident Response Planning (IRP), Security Information and Event Management (SIEM), and Signature-based and Heuristic Detection.

MFA is employed as an additional layer of protection in several sectors to help prevent unauthorized access to sensitive data. IRP is a plan in place to address cybersecurity problems efficiently and expeditiously. SIEM offers real-time analysis and alerts the system of threats and vulnerabilities. Heuristic-based detection relies on detecting anomalies when it comes …

Federated Learning And Applications In Cybersecurity, Ani Sreekumar

Cybersecurity Undergraduate Research Showcase

Machine learning is a subfield of artificial intelligence that focuses on making predictions about some outcome based on information from a dataset. In cybersecurity, machine learning is often used to improve intrusion detection systems and identify trends in data that could indicate an oncoming cyber attack. Data privacy is an extremely important aspect of cybersecurity, and there are many industries that have more demanding laws to ensure the security of user data. Due to these regulations, machine learning algorithms can not be widely utilized in these industries to improve outcomes and accuracy of predictions. However, federated learning is a recent …

A Call For Research: Ethical Dilemmas Of Autonomous Vehicle Manufacturers, Remy Harwood

Cybersecurity Undergraduate Research Showcase

While autonomous vehicles accounted for about 31.4 million vehicles on the road in 2019 (Placek). They have continued to flood the market and have a projected growth to 58 million in just 8 years from now (Placek) As well as a market cap in the billions of dollars. Even the comparatively new AV company Tesla has over 3 times the market cap value of the leading non AV brand Toyota (Market Cap) who are also working toward AVs as well, like their level 2 teammate driver assistance. Following Moore’s Law, as technology continues to improve, their social impact and ethical …

Deep Learning: The Many Approaches Of Intrusion Detection System Can Be Implemented And Improved Upon, Trinity Taylor

Cybersecurity Undergraduate Research Showcase

For my research topic I decided to look at Deep learning. Deep learning can be used in many ways for example in web searching. Deep learning can also can improve new businesses and products. Deep learning could lead to amazing discoveries. Deep learning is making a neural network learn something. In my research I talk about Intrusion detection system, traditional approach for intrusion detection, existing intrusion detection, machine learning and deep learning based intrusion detection system, and future work.

Objective Measure Of Working Memory Capacity Using Eye Movements, James Owens, Gavindya Jayawardena, Yasasi Abeysinghe, Vikas G. Ashok, Sampath Jayarathna

Undergraduate Research Symposium

Human-autonomy teaming (HAT) has become an important area of research due to the autonomous systems being developed for different applications, such as remotely controlled aircraft. Many remotely controlled vehicles will be controlled by automated systems, with a human monitor that may be monitoring multiple vehicles simultaneously. The attention and working memory capacity of operators of remote-controlled vehicles must be maintained at appropriate levels during operation. However, there is currently no direct method of determining working memory capacity, which is important because it is a measure for how memory is being stored for a short term and interacting with long term …

Two-Stage Transfer Learning For Facial Expression Classification In Children, Gregory Hubbard, Megan Witherow, Khan Iftekharuddin

Undergraduate Research Symposium

Studying facial expressions can provide insight into the development of social skills in children and provide support to individuals with developmental disorders. In afflicted individuals, such as children with Autism Spectrum Disorder (ASD), atypical interpretations of facial expressions are well-documented. In computer vision, many popular and state-of-the-art deep learning architectures (VGG16, EfficientNet, ResNet, etc.) are readily available with pre-trained weights for general object recognition. Transfer learning utilizes these pre-trained models to improve generalization on a new task. In this project, transfer learning is implemented to leverage the pretrained model (general object recognition) on facial expression classification. Though this method, the …

Mitigation Of Algorithmic Bias To Improve Ai Fairness, Kathy Wang

Cybersecurity Undergraduate Research Showcase

As artificial intelligence continues to evolve rapidly with emerging innovations, mass-scale digitization could be disrupted due to unfair algorithms with historically biased data. With the rising concerns of algorithmic bias, detecting biases is essential in mitigating and implementing an algorithm that promotes inclusive representation. The spread of ubiquitous artificial intelligence means that improving modeling robustness is at its most crucial point. This paper examines the omnipotence of artificial intelligence and its resulting bias, examples of AI bias in different groups, and a potential framework and mitigation strategies to improve AI fairness and remove AI bias from modeling techniques.

Understanding The Effectivity And Increased Reliance Of Credit Risk Machine Learning Models In Banking, Grishma Baruah

Cybersecurity Undergraduate Research Showcase

Credit risk analysis and making accurate investment and lending decisions has been a challenge for the financial industry for many years, as can be seen with the 2008 financial crisis. However, with the rise of machine learning models and predictive analytics, there has been a shift to increased reliance on technology for determining credit risk. This transition to machine learning comes with both advantages, such as potentially eliminating human error and assumptions from lending decisions, and disadvantages, such as time constraints, data usage inabilities, and lack of understanding nuances in machine learning models. In this paper, I look at four …

Protection Of Patient Privacy On Mobile Device Machine Learning, Matthew Nguyen

Cybersecurity Undergraduate Research Showcase

An existing StudentLife Study mobile dataset was evaluated and organized to be applied to different machine learning methods. Different variables like user activity, exercise, sleep, study space, social, and stress levels are optimized to train a model that could predict user stress level. The different machine learning methods would test if both patient data privacy and training efficiency can be ensured.

Tackling Ai Bias With Gans, Noam Stanislawski

Cybersecurity Undergraduate Research Showcase

Throughout the relatively short history of artificial intelligence (AI), there has been a significant concern surrounding AI’s ability to incorporate and maintain certain characteristics which were not inherently modeled out in its coding. These behaviors stem from the prominent usage of neural network AI, which can inherit human biases from the input data it receives. This paper argues for two possible avenues to combat these biases. The first is to rethink the traditional framework for neural network projects and retool them to be usable by a Generative Adversarial Network (GAN). In a GAN’s zero-sum game, two network techniques can combat …

Fundamentals Of Human-Centric Artificial Intelligence (A.I.): Comparative Analysis Of Europe And The U. S. Landscape, Torré A. Williams

Cybersecurity Undergraduate Research Showcase

This research is a comparative analysis of human-centric Artificial Intelligence (A.I.) in Europe and the U.S. This research establishes fundamentals that are critical to what makes A.I. human-centric. This research contains eight phases: 1) Lawful A.I.; 2) Robust A.I.; 3) Ethical A.I.; 4) Human-centric A.I.; 5) Current State of A.I.; 6) A.I. in Europe; 7) A.I. in the U.S.; 9) Importance of Human-centric A.I. This research shows that there are still ongoing changes with having a human-centric A.I. and why it is very important to society. This research is the beginning of the making of a successful and reliable human-centric …

Data And Artificial Intelligence: Mismatch Between Expectations And Uses, Diana Garcia

Cybersecurity Undergraduate Research Showcase

People like to hide behind their phones when it comes to social media. Not every user has their real name or their own photo on display in their social media account. To obfuscate their identities, some users use unusual usernames and profile photos that are divorced from their true identity.

Accessibility Of Deepfakes, Andrew L. Collings

Cybersecurity Undergraduate Research Showcase

The danger posed by falsified media, commonly referred to as deepfakes, has been well researched and documented. The software Faceswap to was used to swap the faces of two politician (Joe Biden and Donald Trump). The testing was performed using an affordable consumer GPU (an AMD Radeon RX 570) over 100,000 iterations. The process and results for the two attempts with the best results (and largest differences) were recorded. The result was ultimately unconvincing, while the software was able to recreate the facial structure the lighting and skin tone did not blend at all.