Other Computer Engineering Commons^™

Personalized Medical Predictions, Bhargavi Chevva

Master's Projects

Over the past few years, personalized medicine has gained traction due to its ability to solve medical issues efficiently for a person based on their personal characteristics. This project aims to design a machine-learning model that can generate predictions of lab test scores in the future based on past medical history. The model is trained using the MIMIC-4 (Medical Information Mart for Intensive Care) dataset that consists of medical records of over 40,000 patients. The proposed model, MOE-BEHRT, consists of Bidirectional Encoder Representations from Transformers on Electronic Health Records (BEHRT) and Mixture of Experts (MOE). The BEHRT model was originally …

Novel Approach To Music Analysis Using Apache Spark, Nidhi Zare

Master's Projects

Music is one of the most common source of entertainment. Every user has their own taste of music and prefer to listen music that adheres to their taste and mood. There are various categories, called as music genres in which music can be classified. This research project addresses the challenge in music genre classification by using various deep learning models such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Very Deep Convolutional Networks (VGGNet), ResNet and others. The primary objective of this research is to enhance the accuracy of music genre classification using a distributed computing framework Apache Spark. …

Birdsong Classification Using Deep Learning And Mixit, Sasanka Kosuru

Master's Projects

The identification of bird species using deep learning techniques presents a novel approach in bioacoustics, by significantly advancing our understanding and enhancing our capabilities in bird species recognition from audio recordings. The value of audio over visual data for monitoring ecological patterns in birds can be highlighted with the deployment of automated recording devices in remote wildlife sensing, offering a more cost-effective, non-invasive, and practical solution. However, the methods of processing and classifying the audio remain challenging due to the complexity of bird audio, characterized by diverse vocalizations and imminent environmental noise, which poses difficult challenges to perform effective classification. …

Ranking-Based Hashtag Recommendation With Collaborative And Content-Based Filtering, Fei Pan

Master's Projects

The purpose of this project is recommending relevant hashtags for users using both Collaborative Filtering (CF) and Content-based filtering with Twitter dataset. The Twitter dataset was collected by leveraging Twitter API v2. After data preprocessing, 40,806 tweets posted by 278 users with 3,107 hashtags from 01/01/2022 to 04/30/2022 are used for model training and testing. For CF models, we will mainly focus on generating embeddings to learn about user and hashtag latent factors and finally predict a probability for unseen hashtags with most possibility will be ranked as topK items for corresponding users. In this project, Matrix Factorization (MF), Neural …

Optimization Of Inter-Satellite Routing Using Lstm-Based Path Prediction Model, Yash Bhamare

Master's Projects

Satellite networks are one of the most important components that fulfill the world’s need for connectivity. To ensure that communication is efficient and reliable, robust routing algorithms are a must. Because, although it is true that certain routing characteristics may not be permanently and continuously flawless, a routing technique must effectively adapt to modifications in such network characteristics. The new routing method uses a Long Short-Term Memory (LSTM) model to manage dynamic metrics for Low Earth Orbit satellite networks. This LSTM model is aimed at predicting the optimal routing direction on the premise that a satellite is soon to be, …

Explaining The Maliciousness Of Urls Using Shap And Lime, Ayush Nair

Master's Projects

No system has ever reached the levels of proliferation that the Internet now enjoys. It stands as the most widely spread distributed system across the globe; yet this evolution has given rise to an ever-growing wave of malintent that challenges every user and entity on the vast expanse of cyberspace. Malicious URLs loom large as vulnerabilities leaving users naked as they traverse online landscapes, but cybersecurity experts craft models with esoteric algorithms in a bid to stem this tide and shield users from cybercrime. However, peering into the decision-making corridors of these models holds key importance, it’s through understanding such …

Fake Malware Generation Using Gans As Api Calls, Saieswar Reddy Vaka

Master's Projects

With malware threats on the rise, they have also grown more complicated and subtle. Consequently, incorporating cutting-edge machine learning into cybersecurity defenses has never been more crucial. Nevertheless, building resilient machine-learning models is a significant challenge due to the need for existing diversified and complete malware datasets. This project will relieve this difficulty by employing a Generative

Adversarial Network (GAN) to develop artificial malware samples featuring an Appli- cation Programming Interface (API) call series. While traditional generative modeling

has primarily been limited to image-based fields, we offer an “outside the box” domain – malware signature generation – as an API …

Sudoku As A Proof Of Useful Work Protocol On The Blockchain, Abishek Padaki

Master's Projects

The Proof of Work (PoW) consensus used by many blockchain networks like Bitcoin has been criticized for its excessive energy consumption and lack of tangible utility beyond maintaining the network. This report proposes a Proof of Useful Work (uPoW) protocol as an alternative consensus mechanism that utilizes computational resources to solve intrinsically valuable problems. Specifically, it explores the implementation of uPoW on the SpartanGold blockchain test network, where miners must solve Sudoku puzzles to validate new blocks. The report examines the shortcomings of traditional Proof-of-Work protocols, such as their environmental impact and inefficient use of computing power. It then delves …

Reinforcement Learning-Based Dynamic Pricing For Revenue Maximization With Elastic Network Slicing, Jovian Anthony Jaison

Master's Projects

Network slicing is a key enabler of next-generation networking that supports a diverse array of network applications with different service requirements. In particular, elastic network slicing that dynamically scales the bandwidth reserved for each network slice would benefit both slice users and providers through cost-effective resource utilization. However, the elasticity poses a complex problem of managing the dynamics of fluctuating network slices. It is necessary for a slice provider to maintain the balance of different types of slice requests, so it can accommodate more requests while satisfying the service requirements for each slice type. Dynamic pricing of slice resources is …

Ai-Driven Credit Card Fraud Detection With Enhanced User Interaction, Toshi Bhat

Master's Projects

This thesis describes the development and testing of a unique system for detecting credit card fraud. The system employs graph neural networks (GNNs) and a real-time user interaction platform. The primary goal of this study is to use advanced machine learning methods and interactive technologies to improve fraud detection accuracy and the speed with which users can receive assistance. GraphSAGE, a type of GNN, was trained on a simulated set of credit card transactions, allowing the system to detect and predict fraud very accurately. Simulating a real-world transaction scenario is an important aspect of the project. In this case, the …

Nft Price Prediction Using Machine Learning, Akhil Patil Bagili

Master's Projects

In the evolving cryptocurrency marketplace, Non Fungible Tokens (NFT) pose a unique challenge when it comes to predicting the prices due to their high volatility and fluctuating nature. This project aims to create a model that utilizes deep learning techniques to accurately forecast NFT prices. Based on the real time data and the transaction history, the model uses Convolutional Neural Networks (CNNs) and Long Term Short Memory Networks (LSTM) to analyze and make effective predictions about future prices. The methodology involves gathering data from two online marketplaces, Dune and Opensea to create a dataset that enhances the model’s predictive capabilities. …

Base Station Selection Based On The Predicted Dwell Time In 5g-V2x Handover, Anushree Jayesh Shah

Master's Projects

Vehicle-to-Everything (V2X) networks have facilitated smooth communication in vehicles and their surroundings via 5G technology. As vehicles move into different coverage areas, they tend to switch between cellular base stations in order to talk to these "things". Handovers are crucial to maintaining these networks’ effectiveness. To ensure uninterrupted connectivity and seamless handovers, it is extremely important to select a base station that is most appropriate. A vehicle’s dwell time is the duration it stays connected to a base station before handoff. This research focuses on predicting dwell time using regression techniques to proactively choose the most suitable base station. We …

Multiclass Lung Disease Classification In Chest X-Ray Images: A Fine-Tuned Hybrid Cnn-Gnn Approach Using Transfer Learning And Feature Extraction, Vinay Vilas Khade

Master's Projects

In clinical practice, it is still difficult to accurately diagnose lung diseases from chest X-ray (CXR) images. In this study, we propose a new hybrid method for identifying several types of lung diseases using CXR images by combining Convolutional Neural Networks (CNNs) with Graph Neural Networks (GNNs). The framework of our proposed methodology takes advantage of CNN’s ability to extract detailed visual features and GNN’s capacity to understand complex relationships between these features, enabling comprehensive analysis in a multi-class classification setting of COVID-19, pneumonia, and normal lung conditions. We used multiple transfer learning models such as DenseNet201, VGG16, VGG19, MobileNetV2, …

Optimized Community Detection Across Distributed Heterogeneous Servers, Akash Narang

Master's Projects

The exploration of community detection is crucial across various fields, including marketing, and biological research. This area has evolved from non-overlapping communities to recognize nodes as part of multiple overlapping communities. Current research continues to uncover these dynamics. The main challenge is identifying overlapping communities in graphs with billions of nodes and edges. This paper aims to enhance methodologies for community detection in parallel for unprecedentedly large and complex networks. We introduce the HeteroNodesAdapter algorithm, which supports heterogeneous worker nodes and optimized load distribution in graph stream processing. Additionally, we propose the TailBalancedCommunitySize algorithm to find an optimum community size, …

Detecting Fake Reviews Using Aspect Based Sentiment Analysis And Graph Convolutional Networks, Prathana Phukon

Master's Projects

Online reviews significantly influence consumer behavior and business reputa- tions. Detecting fake reviews is important for maintaining trust and integrity in

these platforms. In this project, an application of Aspect-Based Sentiment Analy- sis (ABSA) called FakeDetectionGCN is introduced to distinguish genuine feedback

from deceptive content. The idea is to analyze sentiments related to specific aspects (features) within reviews. Graph Convolutional Networks (GCNs) are used to model the complex contextual dependencies in the review texts. Additionally, SenticNet, an external semantic resource, is integrated to enhance the understanding of sentiments in the reviews. This model is capable of identifying both human-generated as …

Influence Maximization Using Triadic Closures, Communities, And Quotas, Matthew Fu

Master's Projects

Online social networks have exploded in popularity in the last decade. In addition, traditional advertising methods such as television advertising have greatly decreased. This allows companies to utilize viral marketing more effectively. With viral marketing, companies can spread information on a product to a social network by reaching out to a small group of early adopters, who will go on to inform the people around them of the product. The problem is selecting the early adopters that can maximize the spread of influence. The Influence Maximization (IM) problem is finding a social network’s most influential (early adopters) starting nodes, called …

Ensemble Model With Meta-Learning For Ddos Attack Classification In Sdn, Ankith Indrakumar

Master's Projects

In response to the security threats posed by Distributed Denial of Service (DDoS) attacks, this paper presents an intrusion detection framework with a high-accuracy multi-class classification model. In addition to detecting the existence of DDoS attacks, our framework aims to identify the type of attack (e.g., protocol or message type) so that the system can select the most appropriate countermeasure against the DDoS type. We leverage a meta-learner to build an ensemble model of multiple machine learning models such as LSTM, RF, and KNN to enhance detection and classification accuracy. Tested on the CIC-DDoS 2019 dataset, the proposed model achieves …

Opinion Graphs Construction For Reviews Using Transfer Learning And Large Language Models, Yichen Lin

Master's Projects

With the rapid development of the Internet, reading online reviews before making a purchase, booking a hotel, or making a restaurant reservation has become a part of daily life. Customers often consider reviews as crucial supplementary information before making decisions on how to spend their money. However, reading many reviews to gain helpful information takes time and effort. This project proposes a new method OpinionGraphGenerator that aims to create opinion graphs from hotel reviews to reduce the high volume of text in reviews while preserving essential insights. In an opinion graph, vertices are semantically similar opinions, where each opinion consists …

Community Detection Using Deep Learning: Variational Graph Autoencoder Enhanced With Leiden And K-Truss Techniques, Jyotika Hariom Patil

Master's Projects

Community detection in networks is essential for understanding the complex structures of connected systems. Traditional deep learning (DL) methods such as Graph Neural Networks (GNNs) and Graph Convolutional Networks (GCNs) have shown promised results in supervised tasks, like classification, but often fail in unsupervised tasks like community detection because of the lack of labels. Self- supervised approaches where we integrate crucial community information offer a solution. This project seeks to explore DL methods for community detection, focusing specifically on using Graph Variational Autoencoders (VGAEs). While classical approaches can efficiently handle small to medium-sized networks, they typically struggle with larger-sized structures. …

Adaptive Bounded-Confidence Model For Opinion Maximization, Jacob Ortiz

Master's Projects

Social networks have become a significant source of information due to their easy accessibility, low cost, and ability to spread information quickly. Opinions are crucial in shaping our communication and decision-making processes, and our social connections significantly influence them. We model opinions as continuous values from 0 to 1, i.e., 0 means strong disagree and 1 means strong agree. Each agent has an initial opinion as well as a confidence about her opinion and through interactions with the other agents both are updated. Opinion maximization has gained popularity due to social media’s growing impact on our daily lives, where we …