Computer Sciences Commons^™

Fine-Tuned Clip Models Are Efficient Video Learners, Hanoona Rasheed, Muhammad Uzair Khattak, Muhammad Maaz, Salman Khan, Fahad Shahbaz Khan

Computer Vision Faculty Publications

Large-scale multi-modal training with image-text pairs imparts strong generalization to CLIP model. Since training on a similar scale for videos is infeasible, recent approaches focus on the effective transfer of image-based CLIP to the video domain. In this pursuit, new parametric modules are added to learn temporal information and inter-frame relationships which require meticulous design efforts. Furthermore, when the resulting models are learned on videos, they tend to overfit on the given task distribution and lack in generalization aspect. This begs the following question: How to effectively transfer image-level CLIP representations to videos? In this work, we show that a …

Person Image Synthesis Via Denoising Diffusion Model, Ankan Kumar Bhunia, Salman Khan, Hisham Cholakkal, Rao Muhammad Anwer, Jorma Laaksonen, Mubarak Shah, Fahad Shahbaz Khan

Computer Vision Faculty Publications

The pose-guided person image generation task requires synthesizing photorealistic images of humans in arbitrary poses. The existing approaches use generative adversarial networks that do not necessarily maintain realistic textures or need dense correspondences that struggle to handle complex deformations and severe occlusions. In this work, we show how denoising diffusion models can be applied for high-fidelity person image synthesis with strong sample diversity and enhanced mode coverage of the learnt data distribution. Our proposed Person Image Diffusion Model (PIDM) disintegrates the complex transfer problem into a series of simpler forward-backward denoising steps. This helps in learning plausible source-to-target transformation trajectories …

Grm: Generative Relevance Modeling Using Relevance-Aware Sample Estimation For Document Retrieval, Iain Mackie, Ivan Sekulic, Shubham Chatterjee, Jeffrey Dalton, Fabio Crestani

Computer Science Faculty Research & Creative Works

Recent studies show that Generative Relevance Feedback (GRF), using text generated by Large Language Models (LLMs), can enhance the effectiveness of query expansion. However, LLMs can generate irrelevant information that harms retrieval effectiveness. To address this, we propose Generative Relevance Modeling (GRM) that uses Relevance-Aware Sample Estimation (RASE) for more accurate weighting of expansion terms. Specifically, we identify similar real documents for each generated document and use a neural re-ranker to estimate their relevance. Experiments on three standard document ranking benchmarks show that GRM improves MAP by 6-9% and R@1k by 2-4%, surpassing previous methods.

The Power Of (Virtual) Convergence: The Unrealized Potential Of Pair Programming And Remote Work, Mikayla Maki

University Honors Theses

Remote work is expensive. It can lead to isolation, miscommunications, and ossified organizations. These problems occur because of a synchronicity mismatch between how we need to communicate as humans, and what today's tools are capable of. This mismatch can be solved by the adoption of remote pair programming, as exemplified by the authors work at a startup (Zed). Pair programming provides the organic, synchronous, reciprocal interaction necessary to develop the sorts of relationships that remote firms currently lack.

Neutrosophic Framework For Assessment Challenges In Smart Sustainable Cities Based On Iot To Better Manage Energy Resources And Decrease The Urban Environment's Ecological Impact, Samah Ibrahim Abdel Aal

Neutrosophic Systems with Applications

Sustainable smart cities based on the Internet of Things (IoT) technology provide promising prospects for improving quality of life. However, in order to facilitate the widespread implementation of IoT-based smart city solutions, there is a need to concern about data privacy and security, standardization, interoperability, scalability, and sustainability. Reducing the environmental effect of urban activities, optimizing the management of energy resources, and designing novel services and solutions for inhabitants are all examples of how the smart city concept is inextricably linked to sustainability. There is a need to assess challenges in smart sustainable cities based on IoT. This paper is …

Attention Visual, Baris Dingil

College of Computing and Digital Media Dissertations

This research presents an innovative approach to improving visual-spatial attention using a research tool based on the web. Recognizing the significant role visual-spatial attention plays in everyday life and cognitive function for humans, this research was undertaken with the aim of developing a user-friendly, accessible web-based tool called Attention Visual (attentionvisual.com) to enhance this crucial cognitive skill. This tool also facilitates data collection, potentially accelerating the pace and enhancing the quality of related research. Both qualitative and quantitative methods were utilized for data collection and analysis. In order to stimulate improvements in visual-spatial attention, the tool’s algorithm was structured to …

Preliminary Perspectives On Information Passing In The Intelligence Community, Jeremy E. Block, Ilana Bookner, Sharon Lynn Chu, R. Jordan Crouser, Donald R. Honeycutt, Rebecca M. Jonas, Abhishek Kulkarni, Yancy Vance Paredes, Eric D. Ragan

Computer Science: Faculty Publications

Analyst sensemaking research typically focuses on individual or small groups conducting intelligence tasks. This has helped understand information retrieval tasks and how people communicate information. As a part of the grand challenge of the Summer Conference on Applied Data Science (SCADS) to build a system that can generate tailored daily reports (TLDR) for intelligence analysts, we conducted a qualitative interview study with analysts to increase understanding of information passing in the intelligence community. While our results are preliminary, we expect that this work will contribute to a better understanding of the information ecosystem of the intelligence community, how institutional dynamics …

Insect Classification And Explainability From Image Data Via Deep Learning Techniques, Tanvir Hossain Bhuiyan

USF Tampa Graduate Theses and Dissertations

Since the dawn of the Industrial Revolution, humanity has always tried to make labor more efficient and automated, and this trend is only continuing in the modern digital age. With the advent of artificial intelligence (AI) techniques in the latter part of the 20th century, the speed and scale with which AI has been leveraged to automate tasks defy human imagination. Many people deeply entrenched in the technology field are genuinely intrigued and concerned about how AI may change many of the ways in which humans have been living for millennia. Only time will provide the answers. This dissertation is …

Streaming Approximation Scheme For Minimizing Total Completion Time On Parallel Machines Subject To Varying Processing Capacity, Bin Fu, Yumei Huo, Hairong Zhao

Computer Science Faculty Publications

We study the problem of minimizing total completion time on parallel machines subject to varying processing capacity. In this paper, we develop an approximation scheme for the problem under the data stream model where the input data is massive and cannot fit into memory and thus can only be scanned a few times. Our algorithm can compute an approximate value of the optimal total completion time in one pass and output the schedule with the approximate value in two passes.

Job Management Portal Software Review, Ruchir Elukurthy

University Honors Theses

This essay provides an overview of a computer science capstone project focused on developing a website for Abilities At Work, a non-profit organization. The website aims to assist employment specialists in managing clients' information and tracking their job application in finding meaningful employment. The essay highlights the various stages of the project, understanding requirements, selecting tools and technologies, creating an application architecture, and writing code. Also, this essay focuses on the challenges encountered during the project, along with the valuable lessons learned. This essay emphasizes how the project closely resembles real-world software development, offering insights for prospective students and professionals. …

Neutrosophic Framework For Assessment Challenges In Smart Sustainable Cities Based On Iot To Better Manage Energy Resources And Decrease The Urban Environment's Ecological Impact, Samah Ibrahim Abdel Aal

Neutrosophic Systems with Applications

Sustainable smart cities based on the Internet of Things (IoT) technology provide promising prospects for improving quality of life. However, in order to facilitate the widespread implementation of IoT-based smart city solutions, there is a need to concern about data privacy and security, standardization, interoperability, scalability, and sustainability. Reducing the environmental effect of urban activities, optimizing the management of energy resources, and designing novel services and solutions for inhabitants are all examples of how the smart city concept is inextricably linked to sustainability. There is a need to assess challenges in smart sustainable cities based on IoT. This paper is …

Aisha: A Custom Ai Library Chatbot Using The Chatgpt Api, Yrjo Lappalainen, Nikesh Narayanan

All Works

This article focuses on the development of a custom chatbot for Zayed University Library (United Arab Emirates) using Python and the ChatGPT API. The chatbot, named Aisha, was designed to provide quick and efficient reference and support services to students and faculty outside the library's regular operating hours. The article also discusses the benefits of chatbots in academic libraries, and reviews the early literature on ChatGPT's applicability in this field. The article describes the development process, perceived capabilities and limitations of the bot, and plans for further development. This project represents the first fully reported attempt to explore the potential …

A Deep Hierarchical Variational Autoencoder For World Models In Complex Reinforcement Learning Environments, Sriharshitha Ayyalasomayajula

Dissertations and Theses

Model-based reinforcement learning (MBRL) approaches leverage learned models of the environment to plan and make optimal decisions, reducing the need for extensive real-world interactions and enabling more efficient learning in complex domains such as robotics, autonomous systems, and resource allocation problems. They also provide interpretability and insight into the underlying dynamics, facilitating better decision-making and system understanding.

The world model is a model-based RL approach that employs generative neural network models to learn a compressed spatial and temporal representation of the environment. This work explores world models and a simple single-layered RNN model to learn a simple policy based on …

Coordinating Tethered Autonomous Underwater Vehicles Towards Entanglement-Free Navigation, Abhishek Patil, Myoungkuk Park, Jungyun Bae

Michigan Tech Publications, Part 1

This paper proposes an algorithm that provides operational strategies for multiple tethered autonomous underwater vehicle (T-AUV) systems for entanglement-free navigation. T-AUVs can perform underwater tasks under reliable communication and power supply, which is the most substantial benefit of their operation. Thus, if one can overcome the entanglement issues while utilizing multiple tethered vehicles, the potential applications of the system increase including ecosystem exploration, infrastructure inspection, maintenance, search and rescue, underwater construction, and surveillance. In this study, we focus on developing strategies for task allocation, path planning, and scheduling that ensure entanglement-free operations while considering workload balancing among the vehicles. We …

Planning For The Preservation Of Biodiversity And Environmental Assets In Support Of Sustainable Planning In Urban Areas, Case Study: Batu Licin City, Tanah Bumbu Regency, South Kalimantan, Dharma Kalsuma, Hendricus Andy Simarmata

Smart City

Planning for preserving biodiversity and environmental assets in urban areas is crucial in supporting sustainable planning. This study aims to design a strategy for preserving biodiversity and environmental assets in Batu Licin, Tanah Bumbu Regency, South Kalimantan, to support sustainable urban planning. The method used is field observation and literature study. The results showed that the Batu Licin area has a relatively high potential for biodiversity and environmental assets. However, along with the growth of the Batu Licin urban area in the future, it will undoubtedly be a source of pressure on biodiversity. Several strategies can be implemented to overcome …

Say That Again: The Role Of Multimodal Redundancy In Communication And Context, Brandon Javier Dormes

Cognitive Science Senior Theses

With several modes of expression, such as facial expressions, body language, and speech working together to convey meaning, social communication is rich in redundancy. While typically relegated to signal preservation, this study investigates the role of cross-modal redundancies in establishing performance context, focusing on unaided, solo performances. Drawing on information theory, I operationalize redundancy as predictability and use an array of machine learning models to featurize speakers' facial expressions, body poses, movement speeds, acoustic features, and spoken language from 24 TEDTalks and 16 episodes of Comedy Central Stand-Up Presents. This analysis demonstrates that it is possible to distinguish between these …

The Dilemma Of Disclosure: Designing Interpersonal Informatics Tools For Mood Tracking, Daniel Earl Westphal

Computer Science Senior Theses

Mental health is a serious issue that affects people of all ages, but is especially prevalent amongst college age youth. In the 2020-2021 school year, researchers found that around 60% of college students met the criteria for at least one mental health condition, such as major depression or generalized anxiety disorder. Many digital interventions have been innovated in order to help address this issue. These range in type and functionality from teletherapy to medication tracking applications. Some of these digital interventions include social features that allow users to interact with other users, friends, family, or doctors; however, having social features …

How Photorealistic Images Are Generated, Nahom Ketema

University Honors Theses

The field of computer graphics looks into how computers can be used to generate images. From using some trigonometry to plot 3D objects to using rays to calculate the lighting of an object, there are a variety of ways that we can use to draw objects onto a screen. For this thesis, we will be looking at a few of those methods to determine how photorealistic images are generated.

Digital Twin Haptic Robotic Arms: Towards Handshakes In The Metaverse, Mohd Faisal, Fedwa Laamarti, Abdulmotaleb El Saddik

Computer Vision Faculty Publications

More daily interactions are happening in the digital world of the metaverse. Providing individuals with means to perform a handshake during these interactions can enhance the overall user experience. In this paper, we put forward the design and implementation of two right-handed underactuated Digital Twin robotic arms to mediate the physical handshake interaction between two individuals. This allows them to perform a handshake while they are in separate locations. The experimental findings are very promising as our evaluation shows that the participants were highly interested in using our system to shake hands with their loved ones when they are physically …

Epl Card Reader Capstone: The Strengths Of Partner Programming From A Team Leader's Perspective, Zach Yost

University Honors Theses

This essay looks to reflect back upon the successes and failures of the EPL Card Reader capstone project, sponsored by Edward Ivory, head of Portland State University's Electronics Prototyping Lab. The EPL Card Reader's goal is to provide a means of tracking and updating student activity and training on the various machines in the lab. Using a local computer port to host this web app a lab administrator or manager is able to scan a student's access badge to review which machines they have been trained on as well as update that training status. The app also has a running …

On Correspondences Between Feedforward Artificial Neural Networks On Finite Memory Automata And Classes Of Primitive Recursive Functions, Vladimir A. Kulyukin

Computer Science Faculty and Staff Publications

When realized on computational devices with finite quantities of memory, feedforward artificial neural networks and the functions they compute cease being abstract mathematical objects and turn into executable programs generating concrete computations. To differentiate between feedforward artificial neural networks and their functions as abstract mathematical objects and the realizations of these networks and functions on finite memory devices, we introduce the categories of general and actual computabilities and show that there exist correspondences, i.e., bijections, between functions computable by trained feedforward artificial neural networks on finite memory automata and classes of primitive recursive functions.

Utilizing Few-Shot Meta Learning Algorithms For Medical Image Segmentation, Nick Littlefield

Thinking Matters Symposium

Deep learning models can be difficult to train because they require large amounts of data, which we usually do not have or are too expensive to get or annotate. To overcome this problem, we can use few-shot meta-learning, which allows us to train deep learning models with little data. Using a few examples, meta-learning, or learning-to-learn, aims to use the experience learned during training to generalize to unknown tasks. Medical imaging is an industry where it is particularly useful, as there is limited publicly available data due to patient privacy concerns and annotating costs.

This project examines how meta-learning performs …

Joint Flood Risks In The Grand River Watershed, Poornima Unnikrishnan, Kumaraswamy Ponnambalam, Nirupama Agrawal, Fakhri Karray

Machine Learning Faculty Publications

According to the World Meteorological Organization, since 2000, there has been an increase in global flood-related disasters by 134 percent compared to the previous decades. Efficient flood risk management strategies necessitate a holistic approach to evaluating flood vulnerabilities and risks. Catastrophic losses can occur when the peak flow values in the rivers in a basin coincide. Therefore, estimating the joint flood risks in a region is vital, especially when frequent occurrences of extreme events are experienced. This study focuses on estimating the joint flood risks due to river flow extremes in the Grand River watershed in Canada. For this purpose, …

Machine Learning In Action: Exploring Examples In Multiple Domains, German Harvey Alferez

Achieve

Machine learning, a subset of artificial intelligence, is an exciting and fast-moving field in the context of the Fourth Industrial Revolution. Machine learning studies computer algorithms that use a variety of approaches to automatically learn from experience and improve the prediction of a target state, without explicit programming. In this presentation, you will learn about the theoretical foundations of machine learning and how to apply it to solve relevant problems. Special attention will be given to describing the potential of deep learning. This presentation will be illustrated with examples from research projects on geoscience, medicine 4.0, and other areas of …

A Novel Driver Emotion Recognition System Based On Deep Ensemble Classification, Khalid Zaman, Sun Zhaoyun, Babar Shah, Tariq Hussain, Sayyed Mudassar Shah, Farman Ali, Umer Sadiq Khan

All Works

Driver emotion classification is an important topic that can raise awareness of driving habits because many drivers are overconfident and unaware of their bad driving habits. Drivers will acquire insight into their poor driving behaviors and be better able to avoid future accidents if their behavior is automatically identified. In this paper, we use different models such as convolutional neural networks, recurrent neural networks, and multi-layer perceptron classification models to construct an ensemble convolutional neural network-based enhanced driver facial expression recognition model. First, the faces of the drivers are discovered using the faster region-based convolutional neural network (R-CNN) model, which …

The Concept Of Slum Management Become A Tourism Zone In Kendari City, Joko Tri Brata

Smart City

This study aims to assess to find out how the public views the handling of slums in the Kendari City Coastal Area and how the sustainability of slum handling. This research is classified as a qualitative research type, by analyzing the area to be studied because this research is classified as an exploratory research type, by describing the condition and situation of the area that is the object of research. To answer the first research objective related to the pattern of slum handlers, a survey was conducted to find out the pattern of handling slums in Kendari City, carried out …

An Electric Vehicle Analysis Model For Sustainable Environment In Devoicing Nationals, Nada A. Nabeeh, Karam M. Sallam, Ali Wagdy Mohamed

Neutrosophic Systems with Applications

The usage of fossil fuels is regarded as the generation of energy alternative towards electric vehicles (EVs) in third-world nations for a cleaner transportation sector. The rapid development of EVs is the most effective solution, even if the short-term ecological benefits for third-world nations cannot cover the short-term expenses. Since ecological issues have opened the door for certain developing nations to catch up to the worldwide competition, it is important to weigh other options to bring EVs to the marketplace. Hence, the study proposes a model of neutrosophic set combined with entropy to deal with uncertain cases. The proposed model …

Adversary Aware Continual Learning, Muhammad Umer

Theses and Dissertations

Continual learning approaches are useful as they help the model to learn new information (classes) sequentially, while also retaining the previously acquired information (classes). However, these approaches are adversary agnostic, i.e., they do not consider the possibility of malicious attacks. In this dissertation, we have demonstrated that continual learning approaches are extremely vulnerable to the adversarial backdoor attacks, where an intelligent adversary can introduce small amount of misinformation to the model in the form of imperceptible backdoor pattern during training to cause deliberate forgetting of a specific class at test time. We then propose a novel defensive framework to counter …

Machine Learning Data Feature Reduction And Model Optimization, Francisco P. Maturana, Phillip M. Lacasse

AFIT Patents

For machine learning data reduction and model optimization, a method randomly assigns each data feature of a training data set to a plurality of solution groups. Each solution group has no more than a solution group number k of data features and each data feature is assigned to a plurality of solution groups. The method identifies each solution group as a high-quality solution group or a low-quality solution group. The method further calculates data feature scores for each data feature comprising a high bin number and a low bin number. The method determines level data for each data feature from …

Computation Offloading Design For Deep Neural Network Inference On Iot Devices, Asmika Boosarapu

Theses and Dissertations

In recent times, advances in the technologies of Internet-of-Things (IoT) and Deep Neural Networks (DNN) have significantly increased the accuracy and speed of a variety of smart applications. However, one of the barriers to deploying DNN to IoT is the computational limitations of IoT devices as compared with the computationally expensive task of DNN inference. Computation offloading is an approach that addresses this problem by offloading DNN computation tasks to cloud servers. In this thesis we propose a collaborative computation offloading solution, in which some of the work is done on the IoT device, and the remainder of the work …