Computer Sciences Commons^™

Improving Neuropathological Analysis With Aβgan: Addressing Morphology Imbalance For Efficient Alzheimer's Disease Diagnosis, Sujatha Alla, Prasanthi Chidipudi, Nagesh Bheesetty, Vedvikash Reddy Velur, Joshit Mohanty, Puneeth Bheesetty, Marisha Jmukhadze, Narendra Lakshmana Gowda, Sai Gireesh Komaragiri

Engineering Management & Systems Engineering Faculty Publications

Histopathologists are experiencing a digital revolution in their field thanks to the digitization of Whole Slide Images (WSIs), which are microscope slides of tissue that can measure gigapixels in size. With so much high resolution data at their disposal, computer vision techniques can now be used to automate laboratory processes, create visual standards, and increase analysis throughput, all of which reduce the workload of pathologists [1]. The "gold" standard in neuropathology, particularly for Alzheimer's Disease- is pathological diagnosis made by looking at White Matter Inclusions (WSIs) in brain tissue. Semi-quantitative scoring in accordance with the standards established by the Consortium …

Exploring Instructional Designers' Utilization And Perspectives On Generative Ai Tools: A Mixed Methods Study, Tian Luo, Pauline S. Muljana, Xinyue Ren, Dara Young

STEMPS Faculty Publications

The emergence of generative artificial intelligence (GenAI) has caused significant disruptions on a global scale in various workplace settings, including the field of instructional design (ID). Given the paucity of research investigating the impact of GenAI on ID work, we conducted a mixed methods study to understand instructional designers (IDs)’ perceptions and experiences of utilizing GenAI across a spectrum of ID tasks. A total of 70 IDs completed an online survey, and 13 of them participated in the semi-structured interviews. The survey results indicated IDs’ familiarity with and perceived usability of GenAI tools in performing various ID responsibilities in their …

Uncertainty Quantification In Large Language Models Through Convex Hull Analysis, Ferhat Ozgur Catak, Murat Kuzlu

Engineering Technology Faculty Publications

Uncertainty quantification approaches have been more critical in large language models (LLMs), particularly high-risk applications requiring reliable outputs. However, traditional methods for uncertainty quantification, such as probabilistic models and ensemble techniques, face challenges when applied to the complex and high-dimensional nature of LLM-generated outputs. This study proposes a novel geometric approach to uncertainty quantification using convex hull analysis. The proposed method leverages the spatial properties of response embeddings to measure the dispersion and variability of model outputs. The prompts are categorized into three types, i.e., ’easy’, ’moderate’, and ’confusing’, to generate multiple responses using different LLMs at varying temperature settings. …

Enhancing Water Safety: Exploring Recent Technological Approaches For Drowning Detection, Salman Jalalifar, Andrew Belford, Eila Erfani, Amir Razmjou, Rouzbeh Abbassi, Masoud Mohseni-Dargah, Mohsen Asadnia

Research outputs 2022 to 2026

Drowning poses a significant threat, resulting in unexpected injuries and fatalities. To promote water sports activities, it is crucial to develop surveillance systems that enhance safety around pools and waterways. This paper presents an overview of recent advancements in drowning detection, with a specific focus on image processing and sensor-based methods. Furthermore, the potential of artificial intelligence (AI), machine learning algorithms (MLAs), and robotics technology in this field is explored. The review examines the technological challenges, benefits, and drawbacks associated with these approaches. The findings reveal that image processing and sensor-based technologies are the most effective approaches for drowning detection …

Enabling Ai And Robotic Coaches For Physical Rehabilitation Therapy: Iterative Design And Evaluation With Therapists And Post-Stroke Survivors, Min Hun Lee, Daniel Siewiorek, Asim Smailagic, Alexandre Bernardino, Sergi Bermúdez I Badia

Research Collection School Of Computing and Information Systems

Artificial intelligence (AI) and robotic coaches promise the improved engagement of patients on rehabilitation exercises through social interaction. While previous work explored the potential of automatically monitoring exercises for AI and robotic coaches, the deployment of these systems remains a challenge. Previous work described the lack of involving stakeholders to design such functionalities as one of the major causes. In this paper, we present our efforts on eliciting the detailed design specifications on how AI and robotic coaches could interact with and guide patient’s exercises in an effective and acceptable way with four therapists and five post-stroke survivors. Through iterative …

Infrared Ship Segmentation Based On Weakly-Supervised And Semi-Supervised Learning, Isa Ali Ibrahim, Abdallah Namoun, Sami Ullah, Hisham Alasmary, Muhammad Waqas, Iftekhar Ahmad

Research outputs 2022 to 2026

Existing fully-supervised semantic segmentation methods have achieved good performance. However, they all rely on high-quality pixel-level labels. To minimize the annotation costs, weakly-supervised methods or semi-supervised methods are proposed. When such methods are applied to the infrared ship image segmentation, inaccurate object localization occurs, leading to poor segmentation results. In this paper, we propose an infrared ship segmentation (ISS) method based on weakly-supervised and semi-supervised learning, aiming to improve the performance of ISS by combining the advantages of two learning methods. It uses only image-level labels and a minimal number of pixel-level labels to segment different classes of infrared ships. …

Causal Disentangled Recommendation Against User Preference Shifts, Wenjie Wang, Xinyu Lin, Liuhui Wang, Fuli Feng, Yunshan Ma, Tat‑Seng Chua

Research Collection School Of Computing and Information Systems

Recommender systems easily face the issue of user preference shifts. User representations will become outof-date and lead to inappropriate recommendations if user preference has shifted over time. To solve theissue, existing work focuses on learning robust representations or predicting the shifting pattern. Therelacks a comprehensive view to discover the underlying reasons for user preference shifts. To understand thepreference shift, we abstract a causal graph to describe the generation procedure of user interaction sequences.Assuming user preference is stable within a short period, we abstract the interaction sequence as a set ofchronological environments. From the causal graph, we find that the changes …

Societal Impacts Of Artificial Intelligence: Ethics, Legal, And Governance Issues, Yuzhou Qian, Keng Siau, Fiona Fui-Hoon Nah

Research Collection School Of Computing and Information Systems

Artificial intelligence (AI) is quickly changing the way we work and the way we live. The emergence of ChatGPT has thrust AI, especially Generative AI, into the spotlight. The societal impact of AI is on most people's minds. This article presents several research projects on how AI impacts work and society. Three research works are discussed in this article. The first study develops a theoretical framework structuring the legal and ethical objectives that are needed and the means to achieve them. The second study concentrates on bias and discrimination issues embedded in AI applications. It focuses on enhancing the collaboration …

Urban Flood Extent Segmentation And Evaluation From Real-World Surveillance Camera Images Using Deep Convolutional Neural Network, Yidi Wang, Yawen Shen, Behrouz Salahshour, Mecit Cetin, Khan Iftekharuddin, Navid Tahvildari, Guoping Huang, Devin K. Harris, Kwame Ampofo, Jonathan L. Goodall

Civil & Environmental Engineering Faculty Publications

This study explores the use of Deep Convolutional Neural Network (DCNN) for semantic segmentation of flood images. Imagery datasets of urban flooding were used to train two DCNN-based models, and camera images were used to test the application of the models with real-world data. Validation results show that both models extracted flood extent with a mean F1-score over 0.9. The factors that affected the performance included still water surface with specular reflection, wet road surface, and low illumination. In testing, reduced visibility during a storm and raindrops on surveillance cameras were major problems that affected the segmentation of flood extent. …

Multi-Activity Student Knowledge And Behavior Modeling Via Transfer Learning, Siqian Zhao

Electronic Theses & Dissertations (2024 - present)

Online education systems have grown in popularity over the past few years, providing abundant opportunities for students to learn. As the number of students using these systems grows, it promotes the development of the Educational Data Mining (EDM) field, which leverages statistical, machine learning, and data mining technologies to explore large-scale educational data and develop methods to better understand student learning.

In this dissertation, we investigate two essential topics in EDM: Student Knowledge Tracing (KT) and Behavior Modeling (BM). KT aims to quantify and model student knowledge gained from learning activities, while BM focuses on tasks such as modeling student …

Sparse Representation Learning For Temporal Networks, Maxwell Mcneil

Electronic Theses & Dissertations (2024 - present)

Temporal networks arise in many domains including activity of social network users, sensor network readings over time, and time course gene expression within the interaction network of a model organism. Data of this type contains a wealth of prior information such as the connectivity among nodes (e.g., a friendship graph), and prior knowledge of expected temporal patterns (e.g., periodicity). Modeling these temporal and network patterns jointly is essential for state-of-the-art performance in temporal network data analysis and mining. Sparse dictionary encoding is one modeling approach for such underlying patterns. However, most classical approaches consider only one dimension of the data …

Bias-Aware Gaze Uniformity Assessment In Group Images, Omkar Kulkarni

Electronic Theses & Dissertations (2024 - present)

Today, more than 5 billion photos are captured every day, with smartphones generating over 94\% of these images. However, despite advancements in technology, achieving aesthetically pleasing group photos remains challenging, especially when it comes to aligning the direction of everyone’s gaze. While current methods focus on facial features, they often fail to ensure consistent gaze direction. The introduction of the iPhone's Live mode, which captures a 1.5-second video snippet along with still images, complicates the selection of the best key photo due to its subjective nature and a lack of publicly available data, especially during the pandemic.

To address these …

Transformer-Enabled Deep Reinforcement Learning For Coverage Path Planning, Daniel B. Tiu

UNF Graduate Theses and Dissertations

Coverage path planning (CPP) is the problem of covering all points in an environment and is a well-researched topic in robotics due to its sheer practical relevance. This paper investigates such an offline CPP problem where the primary objective is to minimize the path length to achieve complete coverage. Furthermore, the literature suggests that taking turns leads to a higher energy use than going straight. To this end, we design a novel objective function that aims to minimize the number of turns as well. We have proposed a deep reinforcement learning (DRL)-based framework that uses a Transformer model. Unlike state-of-the-art …

Ai-Analyst: An Ai-Assisted Sdlc Analysis Framework For Business Cost Optimization, Nuruzzaman Faruqui, Priyabrata Thatoi, Rohit Choudhary, Ivana Roncevic, Hamed Alqahtani, Iqbal H. Sarker, Shapla Khanam

Research outputs 2022 to 2026

Managing the System Development Lifecycle (SDLC) is a complex task because of its involvement in coordinating diverse activities, stakeholders, and resources while ensuring project goals are met efficiently. The complex nature of the SDLC process leaves plenty of scope for human error, which impacts the overall business cost. This paper introduces AI-Analyst, an AI-assisted framework developed using the transformer-based model with more than 150 million parameters to assist with SDLC management. It minimizes manual effort errors, optimizes resource allocation, and improves decision-making processes, resulting in substantial cost savings. The statistical analysis shows that it saves around 53.33% of costs in …

Short, Full, Best: Analysis Of Different Conference Papers, Miguel Williams

Graduate Research Theses & Dissertations

Research, publish, repeat. This is the basic cycle of anyone in academia. Individuals in academia conduct research, write up your research into a paper or journal article, submit to a conference or journal, and repeat the process. If you're skilled you may even obtain the coveted best paper award from the conference. In this research, I compare full papers to short papers and full papers to best papers. I start by fine-tuning three transformer models for classification capabilities. After I calculate lexical diversity and readability metrics of the papers, I use the features to train three traditional machine learning models. …

Simulation Of A Pick And Place System For Electronic Cards Using A Yumi Cobot, Derrick Sze, Rosula Sj Reyes, Patricia Angela R. Abu

Electronics, Computer, and Communications Engineering Faculty Publications

Collaborative Robots are one of the main drivers of Industry 4.0, which started as a vision focusing on industrial production. It addresses several challenges in the current manufacturing industry such as performing repetitive work and requiring highly skilled workers. The goal of the research is to be able to simulate a pick and place environment with electronic cards using a YuMi cobot and mobile platforms in Coppeliasim. The mobile robot is responsible for transporting the electronic cards to the target location through path planning implemented using the OMPL plug-in. After arriving at the target location, YuMi will then perform the …

"Ai's Pompous Vocabulary Sounds Like It's Trying To Impress Someone": Preservice Teachers' Respond To Ai Reviews Of Their Writing, Jennifer Kidd, Keri Parker, Samantha Myers

Teaching & Learning Faculty Publications

This study examines pre-service teachers’ (PSTs’) responses to using AI to review their writing. As part of a class activity, PSTs prompted an AI-chatbot for feedback and revisions on a short essay based on a provided assignment rubric. Afterward, they reflected on the pros and cons of using AI in this way. While some PSTs initially encountered functionality issues, all were able to obtain feedback and the majority reported finding it useful. In general, the PSTs appreciated AI’s speed and expertise and the potential to customize feedback based on their needs. However, many were skeptical of AI’s evaluation of their …

Poster, Performed: Understanding Public Opinions Of Authorship In Generative Artificial Intelligence Models Via Analogy, Wylie Z. Kasai

Dartmouth College Master’s Theses

Over the last decade, generative artificial intelligence models have advanced significantly and provided the public with several tools to create new works of art. However, the true authorship of these works has been debated due to their training on web-scraped data. Serving as an analogy to these larger models, Poster, Performed is an interactive artificial intelligence exhibition project that uses image assets submitted by the public to create poster compositions with custom image processing algorithms. During the course of a four-day exhibition, visitors were asked to identify the exhibition’s primary artist from five options: (1) participants who submitted image assets, …

A Benchmark Framework For Data Visualization And Explainable Ai (Xai), Murat Kuzlu, Gokcen Ozdemir, Umut Ozdemir

Engineering Technology Faculty Publications

This research introduces a benchmark framework, called EDUMX, designed for machine learning (ML)- based forecasting and XAI tasks, leveraging the Streamlit open-source Python library. The framework offers a comprehensive suite of functionalities, including data loading, feature selection, relationship analysis, data preprocessing, model selection, metric evaluation, training, and real-time monitoring. Users can easily upload data in diverse formats, explore relationships between variables, preprocess data using various techniques, and assess the performance of the ML model using customizable metrics. With its user-friendly interface, this framework offers invaluable insights for forecasting tasks in various domains, catering to the evolving needs of predictive analytics. …

Hack24f: Ai Attacking Ai, Emry Hankins, Sofia Escobar, Yinxin Wan

Paul English Applied Artificial Intelligence (AI) Institute Publications

As MLaaS gains popularity, it also attracts new threats, in particular, model extraction attacks. These attacks involve unauthorized attempts to access and replicate AI models by querying and analyzing the response. Not only do these attacks pose a threat to security and safety but also compromises valuable intellectual property. Because businesses are increasing accessibility to their models that contain sensitive data, it is crucial that they are able to keep them safe and secure.

Hack24f: Careplus: Ai Chatbot, Arianna Fernandez, Sophia Cherkaoui

Paul English Applied Artificial Intelligence (AI) Institute Publications

90% of patients Google symptoms before they talk to their doctor. What are we developing? An app that asks for user symptoms and personal information, users AI to determine medical conditions and provides a plan of action, including recommended medications and suggested doctors to visit.

Hack24f: Llms And World Knowledge Problems, Chengtaon Lin, Joseph Lobo, Huuthanhvy Nguyen, Brian Santos, Ting Xu

Paul English Applied Artificial Intelligence (AI) Institute Publications

While AI “hallucinations” have been widely publicized by the media as shortcomings of Large Language Models (LLMs), there is much less awareness about shortcomings that derive from the potential lack of understanding of the world, from a potential lack of commonsense knowledge. We want to raise awareness among our student body about the potential fallibility of LLMs when they the problems that they need to solve require a basic understanding of the world. We hope that this awareness will prompt everyone to question more carefully the outputs generated by LLMs.

Hack24f: Using Ai To Support Teachers Working With Young Children With Challenging Behaviors, Muqing Zhang, Bennie Bendiksen, Songtian Zeng

Paul English Applied Artificial Intelligence (AI) Institute Publications

While coaching or mental health consultation can be a promising approach to support educators (Kang-Yi et al., 2018), the cost can be substantial and many school districts find it difficult to hire qualified personnel to coach the educators in need. Previous research has attempted to utilize mobile technology to support parents in addressing children with challenging behaviors (Meadan et al., 2016). There are few tools available to support teachers and that attempt to integrate AI technology in the program design and implementation. Our objective is to develop a prototype AI-based tool called Trick-or-Treat to support teachers in identifying and developing …

Hack24f: Beacon Pathway, Michael Agbesi, Ciara Santiago, David Martinez, Olivia Moos, Mohammed Mustafa

Paul English Applied Artificial Intelligence (AI) Institute Publications

The problem: There are a multitude of students from different backgrounds who are feeling lost in their academic journey at UMass Boston & need help with finding opportunities that fit their needs.

The solution: Developing Beacon Pathway, a user-centered platform designed to empower UMass Boston students from diverse backgrounds by streamlining their academic journey. This solution focuses on providing personalized resources, tailored opportunities, and community engagement, ensuring every student feels supported and connected as they navigate their unique paths to success

Hack24f: Tori Project: An Llm Powered Personal Assistant That Speaks, Inal Mashukov, Avi Felzenstein

Paul English Applied Artificial Intelligence (AI) Institute Publications

The problem: remember Jarvis from Iron Man? If not, Jarvis is an AI personal assistant to Tony Stark, aka Iron Man, which can speak in a human-like manner, and is in charge of the systems of the Iron Man suit. Tony Stark also relies on Jarvis for direct access to information, automation of his projects, and helping with work in general. Currently, the “best” real-world alternatives to Jarvis are Siri and Alexa. These are NOT particularly intelligent. Wouldn't it be nice to have a Jarvis-like assistant that is: actually intelligent and resourceful, sounds human, runs locally/offline on your phone or …

Hack24f: Ai And Mental Health: Addressing The Therapy Gap, Samatrai Piam, Digvijay Mahawar, Meric Kinali, Luxman Surentha, Jackson Comeau, Gail Rauch, Cody Turner

Paul English Applied Artificial Intelligence (AI) Institute Publications

Real-World Problem: There is a global shortage of mental health professionals, especially in underserved areas. AI-based chatbots could help fill this gap, but there are serious ethical concerns about the quality of care and efficacy of such chatbots.

Solution: This is a non-technical project that involved conversing with existing chatbots in an effort to develop a normative framework that could be used by future AI therapy bot developers. This framework seeks to build on, and operationalize, some of the insights from the recently published whitepaper on the ethics of chatbot therapy by the UMass Boston Applied Ethics Center. This presentation …

Hack24f: Examine The Human Creativity And Productivity In Ai Generated Content, Jongsoo Ha, Dante Barton, Onkar Khedkar, Neeti Shah, Yash Gondkar, Youxiang Zhu, Xiaohui Liang

Paul English Applied Artificial Intelligence (AI) Institute Publications

This project, presented at the Fall 2024 Hackathon, explores the balance between human creativity and productivity in the age of AI-generated content. With tools such as ChatGPT and Grammarly becoming integral to students’ academic workflows, questions surrounding originality, intellectual property, and academic integrity have grown increasingly urgent. The team addresses the challenge of distinguishing between human and AI contributions in content creation, particularly as current AI detection tools prove unreliable. Their proposed solution involves a system that not only detects AI-generated text using watermarking techniques but also quantifies contributions from humans and AI throughout the writing process.

Identifying And Predicting Patterns Of Snowpack Ripening With Machine Learning Methods, Clement Cherblanc

Graduate Student Theses, Dissertations, & Professional Papers

The timing of water release from the snowpack plays key roles in ecosystem services, groundwater recharge, and water resource management. However, two internal barriers in a standing snowpack must be overcome before runoff can outflow from the base: 1) the cold content must be exhausted, and 2) the interconnected network of snow grains must be filled with liquid water to residual saturation. Expressing the liquid water as latent heat allows the two barriers to be grouped as an energy (J/m²) to define a snowpack’s Runoff Energy Hurdle (REH). The growth and loss of REH is driven by evolution of pore …

Autonomous Strike Uavs In Support Of Homeland Security Missions: Challenges And Preliminary Solutions, Meshari Aljohani, Ravi Mukkamala, Stephan Olariu

Computer Science Faculty Publications

Unmanned Aerial Vehicles (UAVs) are becoming crucial tools in modern homeland security applications, primarily because of their cost-effectiveness, risk reduction, and ability to perform a wider range of activities. This study focuses on the use of autonomous UAVs to conduct, as part of homeland security applications, strike missions against high-value terrorist targets. Owing to developments in ledger technology, smart contracts, and machine learning, activities formerly carried out by professionals or remotely flown UAVs are now feasible. Our study provides the first in-depth analysis of the challenges and preliminary solutions for the successful implementation of an autonomous UAV mission. Specifically, we …

The Combined Focal Loss And Dice Loss Function Improves The Segmentation Of Beta-Sheets In Medium-Resolution Cryo-Electron-Microscopy Density Maps, Yongcheng Mu, Thu Nguyen, Bryan Hawickhorst, Willy Wriggers, Jiangwen Sun, Jing He

Computer Science Faculty Publications

Although multiple neural networks have been proposed for detecting secondary structures from medium-resolution (5–10 Å) cryo-electron microscopy (cryo-EM) maps, the loss functions used in the existing deep learning networks are primarily based on cross-entropy loss, which is known to be sensitive to class imbalances. To monitor and tune the performance of various loss functions for the secondary structure detection problem, we investigated five loss functions: cross-entropy, Focal loss, Dice loss, and two combined loss functions. Using a U-Net architecture in our DeepSSETracer method and a dataset composed of 1,355 box-cropped atomic-structure/density-map pairs, we found that a newly designed loss function …