Session A4: 9:00-10:15, Meade Hall 215, Computer Science/Mathematics and Statistics/Cyber Security and Information Assurance Student Competition Finals
Moderated by Prasanthi Sreekumari
Session A4: 9:00-10:15, Meade Hall 215, Computer Science/Mathematics and Statistics/Cyber Security and Information Assurance Student Competition Finals
Moderated by Prasanthi Sreekumari
(9:00-9:12) Exploring Louisiana Parishes: Broadband, Cancer Rates, and Telehealth
Presented by Poshak Pathak
Poshak Pathak, Prasanthi Sreekumari
Telehealth holds significant importance due to its capacity to deliver remote healthcare services, particularly vital for cancer patients. Improved broadband coverage plays a pivotal role in facilitating these telehealth services. According to the Louisiana Department of Health, the cancer rate in Louisiana has been increasing over the past 40 years. This research focuses on examining broadband coverage in each parish in Louisiana and understanding the cancer rates in those areas. Investigating how broadband availability relates to cancer rates, alongside recognizing the significance of telehealth, allows us to uncover connections that may shed light on the challenges people face in these regions regarding cancer care. The findings from our study could guide efforts to enhance access to resources, healthcare, and telehealth services in these areas.
(9:15-9:27) Construction of a Natural Language Processing Model using Open-Source Text Data
Presented by Dennis Telemacque
Dennis Telemacque
Inspired by advancements like ChatGPT and DALL-E, our research explores the potential of Natural Language Processing (NLP) for text generation. We built a character-level model with a neural network architecture, aiming to optimize its performance for generating grammatically accurate text.
We evaluated various parameters like layers, learning rate, and training epochs, utilizing diverse datasets to train the model. While initial accuracy was commendable, achieving perfect syntactic structures remained a challenge.
Crucially, incorporating a garbage collector and using a specific batch size range significantly improved training and testing accuracy (58% to 80%). This research opens doors for further exploration, particularly in identifying the minimal dataset required for coherent text generation, paving the way for more efficient and effective NLP models
(9:30-9:42) A Detailed Analysis of Data Hiding Approaches for Securing Sensitive Information
Presented by Utsab Neupane
Utsab Neupane, Prasanthi Sreekumari
In a world where digital technology is rapidly advancing, it's crucial to protect sensitive information. This research explores various methods of data concealment and assesses their effectiveness in ensuring data security. By closely examining different data hiding techniques, we gain insights into their strengths and weaknesses, with a focus on their capacity to conceal data, evade detection, and withstand attacks. The primary objective of this research is to assist readers in understanding diverse data hiding techniques, enabling them to choose the most effective way to safeguard their important information in today's ever-changing digital environment.
(9:45-9:57) Unveiling Brain Biometrics: Exploring EEG Pattern Similarities for Identity Verification Using Machine Learning algorithms.
Presented by Emeka Kelvin Iphy
Emeka Kelvin Iphy, Omer Soysal
Over the years, biometric authentication methods such as fingerprint, iris scanning, and facial recognition have improved security in various sectors, including border control, banking, and access control. With the rise of technological power, building a system that permits brain biometrics will add a secure layer of security. This research aims to show whether EEG-based representation helps identify individuals by exploring the uniqueness and permanence of their brain signals.
We present a preliminary result using eight descriptive statistical features that provide information about the distribution of the EEG channels. Previous results using entropy calculation and clustering algorithm show higher overlap in both intra and inter-subjects, which corresponds to higher permanence and lower uniqueness. This affirms stability but not distinctiveness in individuals. We aim to compare different machine learning algorithms, such as Artificial Neural Networks (ANN), Support Vector Machines (SVM), and Decision Trees (DT), to further explore their uniqueness and permanence.
(10:00-10:02) Influence of TikTok on Gen Z
Presented by Gabrielle Johnson
Gabrielle Johnson, Lynette Jackson, Phyllis Okwan
This project studies the influence of TikTok on the societal and governmental views of the Gen Z population. TikTok is an app that consists of short videos that can be anywhere from 10 seconds to 3 minutes. The videos can range from cute dance videos, tutorials, and comedy, to politics, TikTok can be used to research topics and review things. This research examines the platform’s role in shaping perspectives, creating discourse, and viewing a shift in how the younger generation comprehends and interacts with the rest of the world. The research examined how TikTok influences the Gen Z generation with their view on politics and society, for example, a Gen Z can be in one country and can see another person on TikTok live in another country. Gen Z can also use TikTok as their version of the news to keep up with social trends such as petitions, protests, boycotting as they do currently, and also what’s going on in other countries outside of their own if they have any social issues or unrest. A poll was conducted on the extent TikTok can influence views on societal and government issues. The results indicated that TikTok helped influence Gen Z on societal and governmental issues.