Computer Engineering Commons^™

Bridging Human And Machine Intelligence: Reverse-Engineering Radiologist Intentions For Clinical Trust And Adoption, Akash Awasthi, Ngan Le, Zhigang Deng, Rishi Agrawal, Carol C. Hu, Hien Van Nguyen

Computer Science and Computer Engineering Faculty Publications and Presentations

In the rapidly evolving landscape of medical imaging, the integration of artificial intelligence (AI) with clinical expertise offers unprecedented opportunities to enhance diagnostic precision and accuracy. Yet, the "black box" nature of AI models often limits their integration into clinical practice, where transparency and interpretability are important. This paper presents a novel system leveraging the Large Multimodal Model (LMM) to bridge the gap between AI predictions and the cognitive processes of radiologists. This system consists of two core modules, Temporally Grounded Intention Detection (TGID) and Region Extraction (RE). The TGID module predicts the radiologist's intentions by analyzing eye gaze fixation …

Visual Causal Question And Answering With Knowledge Graph Link Prediction, Utkarshani Jaimini, Cory Henson, Amit Sheth

Publications

The ability to answer causal questions is important for any system that requires robust scene under- standing. In this demonstration, we develop a prototype system that leverages our causal link prediction framework, CausalLP. CausalLP framework uses a visual causal knowledge graph and associated knowledge graph embedding for two visual causal question and answering tasks- (i) causal explanation and (ii) causal prediction. In the live demonstration sessions, the participants will be invited to test the efficiency and effectiveness of the system for visual causal question and answering.

Causal Neuro-Symbolic Ai For Root Cause Analysis In Smart Manufacturing, Utkarshani Jaimini, Cory Henson, Amit Sheth

Publications

Root cause analysis is the process of investigating the cause of a failure and providing measures to prevent future failures. It is an active area of research due to the complexities in manufacturing production lines and the vast amount of data that requires manual inspection. We present a combined approach of causal neuro-symbolic AI for root cause analysis to identify failures in smart manufacturing production lines. We have used data from an industry-grade rocket assembly line and a simulation package to demonstrate the effectiveness and relevance of our approach.

Ontology Design Metapattern For Relationtype Role Composition, Utkarshani Jaimini, Ruwan Wickramarachchi, Cory Henson, Amit Sheth

Publications

RelationType is a metapattern that specifies a property in a knowledge graph that directly links the head of a triple with the type of the tail. This metapattern is useful for knowledge graph link prediction tasks, specifically when one wants to predict the type of a linked entity rather than the entity instance itself. The RelationType metapattern serves as a template for future extensions of an ontology with more fine-grained domain information.

Causal Knowledge Graph For Scene Understanding In Autonomous Driving, Utkarshani Jaimini, Cory Henson, Amit Sheth

Publications

The current approaches to autonomous driving focus on learning from observation or simulated data. These approaches are based on correlations rather than causation. For safety-critical applications, like autonomous driving, it’s important to represent causal dependencies among variables in addition to the domain knowledge expressed in a knowledge graph. This will allow for a better understanding of causation during scenarios that have not been observed, such as malfunctions or accidents. The causal knowledge graph, coupled with domain knowledge, demonstrates how autonomous driving scenes can be represented, learned, and explained using counterfactual and intervention reasoning to infer and understand the behavior of …

Resource Harvesting For Parallel Functions In Serverless Workflows, Peiman Fotouhi

LSU Master's Theses

In the rapidly evolving landscape of cloud computing, serverless architectures have gained attention for their scalability and cost-effectiveness. This thesis aims to introduce a novel approach to maximize resource utilization in serverless environments through the concept of harvesting idle resources within Directed Acyclic Graph (DAG)-based workloads. Our proposed solution targets resource harvesting at parallel stages by utilizing Machine Learning models to accurately harvest or accelerate serverless functions. Additionally, we present a scheduling algorithm specifically designed to address the unique requirements of DAG workloads in cloud environments.

The framework leverages dynamic resource allocation techniques to identify and exploit idle resources within …

Performance Analysis Of C++ Parallel Algorithms In Hpx, Srinivas Yadav Singanaboina

LSU Master's Theses

The exponential growth in computational power and the increasing demand for high-performance applications have driven the need for greater parallel efficiency. Over the years, the number of cores in consumer-level CPUs and high-performance computing (HPC) systems has grown significantly. In response, numerous parallel programming li- braries have been developed. Each of these libraries offers unique mechanisms to enhance parallel performance. In this paper, we investigate the performance of five such paral- lel programming backends: C++ std::execution::par, OpenMP, TBB, Taskflow, and HPX. We evaluate these libraries using two sets of benchmarks. The first set focuses on standard C++ STL algorithms, including …

Detecting Data Poisoning Attacks In Federated Learning For Healthcare Applications Using Deep Learning, Mohammed Aljanabi, Sahar Yousif Mohammed, Alaa Hamza Omran

Iraqi Journal for Computer Science and Mathematics

This work introduces a new approach to protecting the data in the healthcare applications of federated learning based on the classification of skin cancer. The recommended solution established and prevents the data poisoning attacks by using deep learning and CNN architectures namely VGG16. In a federated learning system which comprises of ten healthcare facilities, the approach enables the training of models in a collaborative way without compromising the medical data or the patients’ information. Data is meticulously prepared and preprocessed using the Skin Cancer MNIST: According to the HAM10000 dataset. As for the federated learning approach, VGG16’s feature extraction capability …

Vulnerability Discovery On Broadcom Devices With Qemu, Christian Shane Douglas Little

USF Tampa Graduate Theses and Dissertations

System-on-chip (SoC) devices in the form of routers, access points, and smart appliances have become an increasingly prevalent aspect of our infrastructure. Performing vulnerability research on these devices may be exceptionally difficult given the on-hardware limitations, inability to guarantee certain aspects of the system, and lack of information when the appliance’s source code is not available.In this thesis, we present Broadfeather, a QEMU-based emulation target, and framework for dynamic analysis and instrumentation. While Broadfeather supports data types shared across many chips under the Broadcom umbrella, it is primarily geared towards the BCM6755 Wi-Fi 6E capable chipset. Broadfeather also features the …

Prototyping Interactive Tactile Digital Logic Simulations: A Hybrid Approach, Logan Bateman

MS in Computer Science Project Reports

Tactile exhibits are common in museums and on the walls of university halls. However, few (if any) tools exist for creating tactile exhibits for teaching digital logic or computing concepts. This project implemented a framework for creating tactile digital logic simulation exhibits, with a focus on rapid prototyping and distributed architecture. Prototyping allows for fast iteration, with the ability to simulate unlimited hardware components such as buttons, light emitting diodes (LEDs), and other input or output devices. Through the abstraction of implementations and a distributed communication protocol, switching to real hardware is seamless and works in tandem with simulated hardware. …

Enhanced Shoulder-Surfing Cued Recall Graphical Password System: Sequential Passpoint, Titus D. Fofung

Cybersecurity Graduate Research Symposium

During the past two decades, many graphical passwords have been used widely as an alternative to text-based passwords. However, most graphical password systems are plagued by shoulder-surfing problems, usability, and remembering capability. This study proposed a new graphical password called SPP (Sequential PassPoint), allowing users to remember three click-points on two images in specified order and image order. When the image order changes, the click order is reversed. Two decoy images for three random clicks were introduced to enhance the security of SPP. The proposed SPP system was validated both theoretically and empirically

Experimental Study To Assess The Role Of Environment And Device Type On The Success Of Social Engineering Attacks: The Case Of Judgment Errors, Tommy Pollock

Cybersecurity Graduate Research Symposium

No abstract provided.

Assessing Organizational Investments In Cybersecurity And Financial Performance Before And After Data Breach Incidents Of Cloud Saas Platforms, Munther B. Ghazawneh

Cybersecurity Graduate Research Symposium

No abstract provided.

Examining Consumers' Selective Information Privacy Disclosure Behaviors In An Organization's Secure E-Commerce Systems, Patrick I. Offor

Cybersecurity Graduate Research Symposium

No abstract provided.

Dcai: The 4th International Workshop On Data-Centric Ai, Yanjie Fu, Kunpeng Liu, Dongjie Wang

Computer Science Faculty Publications and Presentations

Machine learning traditionally emphasizes developing models for given datasets, but real-world data is often messy, making model improvement insufficient for enhancing performance. Data-Centric AI (DCAI) is an emerging field that systematically improves datasets, leading to significant practical ML advancements. While experienced data scientists have manually refined datasets through trial-and-error and intuition, DCAI approaches data enhancement as a systematic engineering discipline. DCAI represents a shift from focusing on models to the underlying data used for training and evaluation. Despite the dominance of common model architectures and predictable scaling rules, building and using datasets remain labor-intensive and costly, lacking infrastructure and best …

Interpreting Black-Box Time Series Classifiers Using Parameterised Event Primitives, Ephrem Tibebe Mekonnen, Luca Longo, Pierpaolo Dondio

Conference papers

Amidst the remarkable performance of deep learning models in time series classification, there is a pressing demand for methods that unveil their prediction rationale. Existing feature importance techniques often neglect the temporal nature of time series data, focusing solely on segment importance. Addressing this gap, this paper introduces a local model-agnostic method akin to LIME, which generates neighbouring samples by randomly perturbing segments of the original instance. Subsequently, weights are computed for each neighbouring instance based on its distance from the original, elucidating its influence. Parameterised event primitives (PEPs) are then extracted from these perturbed samples, encompassing increasing and decreasing …

Real-Time Detection Of Sea Turtles Using Uav And Neural Networks On Edge Devices, Jose A. Gonzalez Nunez, Jose G. Gonzalez Nunez, Mustafa I. Akbas, Patrick Currier, Nickolas D. Macchiarella

Journal of Aviation/Aerospace Education & Research

Sea turtle populations continue to diminish around the globe for various reasons. Therefore, the need for innovative solutions to monitor sea turtles has been increasing. This research paper focuses on an innovative application of artificial intelligence (AI) and machine learning (ML) together with unmanned aerial vehicles (UAV) to improve sea turtle conservation efforts. We outline the design, implementation, and evaluation of a system that deploys UAVs equipped with high-resolution cameras, coupled with a purpose-built neural network to recognize, classify, and monitor sea turtles. This project thus serves as a platform for understanding the wider applicability and limitations of this technology …

Three-Dimensional Outdoor Object Detection In Quadrupedal Robots For Surveillance Navigations, Muhammad Hassan Tanveer, Zainab Fatima, Hira Mariam, Tanazzah Rehman, Razvan Cristian Voicu

Faculty Articles

Quadrupedal robots are confronted with the intricate challenge of navigating dynamic environments fraught with diverse and unpredictable scenarios. Effectively identifying and responding to obstacles is paramount for ensuring safe and reliable navigation. This paper introduces a pioneering method for 3D object detection, termed viewpoint feature histograms, which leverages the established paradigm of 2D detection in projection. By translating 2D bounding boxes into 3D object proposals, this approach not only enables the reuse of existing 2D detectors but also significantly increases the performance with less computation required, allowing for real-time detection. Our method is versatile, targeting both bird’s eye view objects …

A Plugin-Based Unreal Engine Adapter For Hla-Based Distributed Simulation, Mei Yang, Peng Wang

Journal of System Simulation

Abstract: With the wide application of game engine-based simulation in transportation, military and other fields, the demand for interoperability between game engine and traditional simulations is becoming increasingly strong. For the HLA-based integration of Unreal Engine and the traditional simulations, a plugin-based Unreal Engine adapter for distributed simulation is designed, which enables the rapid development of Unreal Engine federate and the efficient integration. The simulation shows the feasibility of the plugin-based Unreal Engine adapter.

Research On Sequential Design Methods For Satellite Combat Simulation Tests, Yanlin Wang, Zhijun Cheng, Zichen Wang, Jian Zhong

Journal of System Simulation

Abstract: Aiming at the problem that satellite monitoring mission simulation tests cannot take into account the number of sample points and model accuracy in the complex test space, a hybrid sequential test design method for satellite simulation tests based on sample density and local nonlinearity is proposed. Voronoi division is used to describe the density of discrete point distribution, and the nonlinearity is measured with the help of Taylor expansion and sample point neighborhood gradient information. The two are combined to calculate the hybrid metrics, and the sample points are ranked and new ones are added until the stopping criterion …