News & Press Releases

Title: Optimizing Datasets for Lyme Disease Detection Advisor: Iren Valova PhD, Associate Dean - College of Engineering - Professor, Computer & Information Science - University of Massachusetts Dartmouth Committee: Gokhan Kul PhD, Computer & Information Science - University of Massachusetts Dartmouth Firas Khatib PhD, Computer & Information Science - University of Massachusetts Dartmouth Date: Jan 17, 2025 Time: 1pm Location: Zoom https://umassd.zoom.us/j/98403102776?pwd=VKmd3RikQZbqdTkhOaIhoJdyXQE91k.1 Abstract: This thesis focuses on optimizing image datasets through augmentation methods for the detection of Lyme disease. Lyme disease often is accompanied by an erythema migrans rash, but other sorts of rashes could look similar to it. Using a public crowdsourced dataset, the object is to improve the accuracy of YoloV7 through image enhancements and augmentations. The study utilizes a combination of data preprocessing techniques, including CLAHE, photometric deformation, elastic deformation, and mixup to improve image quality and address dataset imbalances. YoloV7, an object detection model was trained on the enhanced dataset to accurately differentiate Lyme-related rashes from other dermatological conditions. The results favored the CLAHE result over the others. This work contributes to the development of more reliable, automated diagnostic tools for individual user. For further information contact Dr. Iren Valova at ivalova@umassd.edu

See description for location

Date: Tuesday, January 21, 2025 Time: 10am Topic: Statistical and Machine Learning Models for System Reliability and Resilience Location: TXT 105 (CSCDR) Abstract: Software reliability and resilience are essential for ensuring dependable system performance, particularly in the face of evolving demands and unexpected disruptions. Traditional reliability models, such as the Non-Homogeneous Poisson Process (NHPP), have been widely used to predict defect occurrence based on testing time or effort. However, these models often fall short of capturing the complexities of real-world systems. Resilience engineering, which focuses on a system's ability to respond to and recover from shocks, has gained significant attention as a complementary approach to reliability. While statistical models provide foundational insights, their rigid assumptions limit flexibility and fail to account for dynamic patterns in defect occurrence and recovery processes. On the other hand, machine learning methods, such as neural networks, offer the potential to model intricate dependencies and non-linear trends. However, these models often require extensive data, which is not always available in resilience engineering contexts, and may lack robustness in long-term predictions. This gap highlights the need for integrated approaches that effectively address the challenges of modeling resilience in systems experiencing varying types and intensities of shocks. In order to address these challenges, this dissertation proposes hybrid approaches including: (i) For defect prediction, we use recurrent neural networks (RNNs) and long short-term memory (LSTM) networks that incorporate covariates to improve predictive accuracy by reflecting key factors driving defect discovery, (ii) To enhance defect classification, we apply Locally Linear Embedding (LLE) as a preprocessing step, which simplifies complex data, allowing classifiers to better interpret defect patterns, and (iii) For resilience assessment, we introduce a transfer function model, a flexible time series approach that considers multiple stressors and recovery patterns. This model captures the dynamic response of a system under varied shocks, allowing for a more accurate resilience evaluation without needing extensive data. By combining machine learning and statistical methods, this dissertation aims to advance both reliability and resilience assessment in software systems, providing robust, adaptable models capable of predicting defects and tracking recovery under complex conditions. These contributions support the development of systems that not only maintain performance but also adapt to future challenges with resilience. ADVISOR(S): Dr. Lance Fiondella, Department of Electrical & Computer Engineering (lfiondella@umassd.edu) COMMITTEE MEMBERS: Dr. Alfa Heryudono, Department of Mathematics Dr. Ruolin Zhou, Department of Electrical & Computer Engineering Dr. Hong Liu, Department of Electrical & Computer Engineering NOTE: All EAS Students are ENCOURAGED to attend.

Textiles Building 105

Welcome back commuters! Please join us at the TV Pit in the Campus Center for our January Commuter Pit Stop, from 10am-2pm. If you have questions please e-mail commuters@umassd.edu - we'd love to hear from you!

Campus Center TV Pit

College of Engineering EAS Doctoral Proposal Defense by Hieu X Ngo (Henry) Date: Wednesday, January 22, 2025 Time: 9:30 AM - 11:30 AM Title: iGenMIF2V2: Intelligent generative fuzzy behavior pattern recognition with visualization-aided validation for digital trials in a federated framework Location: Zoom https://umassd.zoom.us/j/95683754446?pwd=WW1paGw5Q29mdXBMb0E3N3dkUTZ2Zz09 Meeting ID: 956 8375 4446 Passcode: 408420 Advisor: Dr. Julia Hua Fang, Department of Computer and Information Science Committee Members: Dr. Long Jiao, Department of Computer and Information Science Dr. Honggang Wang, Department of Electrical & Computer Engineering Dr. Gang Zhou, Department of Computer Science, William & Mary Abstract Artificial intelligence (AI)-based behavior pattern recognition and machine learning-based privacy-preserving frameworks hold significant promises for advancing digital behavioral trials. However, these advancements face well-known methodological obstacles particularly in multi-site longitudinal studies of human behaviors and their associated disease outcomes. Soft clustering methods such as fuzzy clustering can accommodate actual behavioral changes over time by allowing subjects to have varying degrees of membership across clusters, thus capturing overlapping behavioral trajectory patterns. To address incomplete data arising from evolving behaviors, multiple imputation-based fuzzy clustering can effectively handle different missing data mechanisms in learning high-dimensional, non-normal longitudinal datasets. Despite these capabilities, a computational method to intelligently optimize the fuzzifier, which intertwines fuzzy pattern recognition, validation, and visualization across diverse trials, has yet to be developed. Furthermore, multi-site digital trials are constrained by privacy regulations such as HIPAA, where even full anonymization may not fully prevent re-identification. Federated learning addresses this by sharing model parameters instead of raw data across sites. Therefore, integrating intelligent fuzzy behavior recognition within a federated framework is critical for advancing digital behavioral research. This proposal attempts to enable intelligent generative fuzzy behavior recognition with visualization-aided validation in a federated framework (iGenMIF2V2), designed to resolve theoretical and technical challenges in longitudinal digital trials. The proposed AI algorithm aims to streamline fuzzifier optimization, fuzzy cluster validation, and visualization for incomplete longitudinal digital trial data. The algorithm will use weighted rank aggregation and 3D visualization to intelligently integrate fuzzifier selection and adaptation with visualization-aided cluster validation, capturing complex behavioral trajectories in longitudinal trials. It will also be adaptable to a broad range of fuzzy clustering methods for digital trials. The algorithm will be validated through ablation studies using both real-world and synthetic datasets. Empirical data will include harmonized longitudinal dietary quality data from four randomized controlled trials in Massachusetts. Synthetic datasets will be generated using a fully crossed design with varying parameters. Additionally, the algorithm will be tested on national longitudinal dietary quality studies, comprising over 3.3 million observations from the U.S. To maximize its impact, a generative clustering model based on iGenMIF2V2 will synthesize chronic disease risk groups characterized by varying dietary quality trajectories, demographics, socioeconomic factors, and biomarkers. The final AI algorithm will be implemented online and made publicly accessible. For further information please contact Dr. Julia Hua Fang at hfang2@umassd.edu

See description for location

Lunch and learn about The Janes, the legendary underground abortion networks. For more information, visit the documentary's website: https://www.hbo.com/movies/the-janes

Center for Women, Gender & Sexuality

Ready to spruce up your resume as we start the spring semester? Join our Career Center team at a workshop-style resume program. Bring your resume and be ready to make edits and changes as a career advisor walks through each section of a professional resume. There are two one-hour sessions, so you can sign up for the one that fits your schedule!

MacLean Campus Center, Conference Room, Room 201