Identifying factors that influence undergraduate students’ performance in an introductory programming (CS1) course can enable educators to optimize student success. It has become increasingly important for educators to understand and provide customized aid to their students in formal learning environments. In fast-paced and large-class environments, educators face challenges in identifying students’ learning needs based on diverse demographic and academic profiles. This study investigates how Gender, Prior Programming Experience (PPE), and Grade Point Average (GPA) impact student success in CS1 courses. The dataset consists of 836 students from six semesters (Spring 2021 to Fall 2023), with demographic and prior experience information collected through surveys. Although in the past, researchers have predicted student success using interaction with course material, previous exam scores, and a combination of many other factors, there is a gap in the literature for using specifically the combination of the pre-course factors of PPE, GPA, and gender to predict performance. Using correlation analysis, logistic regression, and Chi-Square tests, we explored the relationships between these factors and student performance, particularly in predicting whether a student would achieve above or below 80% in course exams before the student starts the course. The logistic regression model achieved a 76% accuracy, with higher precision and recall for identifying students scoring below 80%. GPA showed the strongest positive correlation with performance, while PPE and Gender also exhibited statistically significant relationships, though Gender's impact was minimal. These findings suggest that GPA and PPE are useful predictors for early identification of students at risk of under performance, helping educators develop targeted strategies to support students in programming courses.

Authored by
Amanda Nicole Smith (University of Florida), Sage Bachus (University of Florida), and Ashish Aggarwal (University of Florida)

Magnetic resonance imaging (MRI) enables non-invasive, high-resolution analysis of muscle structures. However, automated segmentation remains limited by high computational costs, reliance on large training datasets, and reduced accuracy in segmenting smaller muscles. Convolutional neural network (CNN)-based methods, while powerful, often suffer from substantial computational overhead, limited generalizability, and poor interpretability across diverse populations. This study proposes a training-free segmentation approach based on keypoint tracking, which integrates keypoint selection with Lucas-Kanade optical flow. The proposed method achieves a mean Dice similarity coefficient (DSC) ranging from 0.6 to 0.7, depending on the keypoint selection strategy, performing comparably to state-of-the-art CNN-based models while substantially reducing computational demands and enhancing interpretability. This scalable framework presents a robust and explainable alternative for muscle segmentation in clinical and research applications.

Authored by
Mengyuan Liu (Northeastern University) and Dr. Jeongkyu Lee (Northeastern University)

Machine learning (ML) algorithms and artificial intelligence (AI) systems have already had an immense impact on our society. Lately, AI/ML has shown to be able to create machine cognition comparable to or even better than human cognition for some applications. For emerging applications, AI is also regarded to provide cybersecurity solutions (i.e., AI for cybersecurity) by detecting anomalies, adapting security parameters based on ongoing cyberattacks, and reacting in real-time to combat cyber-adversaries. However, ML algorithms and AI systems are vulnerable to manipulation of data or learning models, biases, and low credible information due to flaws in learning models and input data. Therefore, ML algorithms and AI systems need robust security and correctness (i.e., cybersecurity for AI) to permit fair and trustworthy AI systems. Unfortunately, AI and cybersecurity have traditionally been treated as separate fields, with little emphasis on their intersection in education. The primary goal of this paper is to discover, explore, develop and integrate a scalable instructional approach for integrated AI and cybersecurity (DARE-AI for short) in undergraduate and graduate curricula. This is accomplished by creating a “learning by doing” approach to address emerging AI and cybersecurity issues that are not covered in an integrated way, if at all, in traditional curricula. The experiments are designed to study fair and trustworthy AI, adaptive intrusion detection, online learning, federated learning, distributed learning, and adversarial learning. We present learning outcomes and results using surveys and assessments. The developed DARE-AI modules help train the next-generation STEM workforce with knowledge of integrated cybersecurity and AI that is expected to help not only to meet evolving demands of the US government and industries, but also to improve the nation’s economic security and preparedness.

Authored by
Utsab Khakurel (Howard University) and Prof. Danda B Rawat (Howard University)

This qualitative study investigates the factors influencing Korean international students’ decisions to major in CS (Computer Science) and CE (Computer Engineering) at a U.S. university, as well as their academic experiences. Through semi-structured interviews with two female and two male Korean undergraduates, three core themes emerged: strong personal interest in computing, influences of parental beliefs and societal norms limiting female students (particularly affecting female students), and utility of major for future career. Findings show that while personal interest is a primary motivator, gender biases and parental beliefs can redirect female students from other STEM fields to CS. Additionally, participants cited dissatisfaction with Korea’s public education system, and freedom and flexibility with U.S. education as reasons for studying abroad. The study underscores the need for reforms in K-12 computer and career education and for addressing gender biases to support informed major and career choices among students.

Authored by
Hyeree Cho (Affiliation unknown) and Woongbin Park (Purdue University at West Lafayette)

Cognitive workload assessment and management are critical in managing work efficiency in high-stress environments and long-duration tasks, such as critical infrastructure operations, first-responder responses, healthcare, military, and transportation. A major challenge in developing cognitive assessment algorithms lies in designing an experimental testbed that integrates diverse systems like brain-computer interfaces, physiological sensors, and task-specific hardware for synchronized multi-modal data collection. This paper presents a novel reconfigurable testbed for assessing cognitive workload using Letter N-back, Flanker N-back, and multiple object tracking (MOT) tasks. The testbed features customizable parameters such as trial length, difficulty level, and task complexity, allowing simulation of various stress levels. The integration of Neuroelectrics EEG headsets and Bluetooth-enabled physiological sensors ensures real-time multimodal data acquisition. In addition, the modular design supports future expansion for new tasks and devices, fostering advancements in cognitive neuroscience and human performance research.

Authored by
Yug Patel (Missouri University of Science and Technology), Sanjana Shangle (University of Texas at Dallas), Asir Abrar (Missouri University of Science and Technology), Prof. Venkata Sriram Siddhardh Nadendla (Missouri University of Science and Technology), and Dr. K Krishnamurthy (Missouri University of Science and Technology)

A significant challenge with non-volatile memory (NVM) is the uneven wear resulting from frequent write operations concentrated on specific memory locations. This problem can drastically reduce the memory's lifespan and compromise system reliability. Frequent writes to certain cells accelerate wear-out, causing errors and degraded performance. Managing this wear in NVM systems is crucial to ensuring their long-term efficiency and usability. In this research, we propose a novel solution called Predictive Memory Wear Balancing (PMWB) to address these challenges. PMWB effectively mitigates the wear imbalance problem. Moreover, PMWB integrates dynamic approximation adjustments, making it one of the first approaches to combine wear balancing with adaptive approximation techniques, which further extend memory endurance without significantly affecting system performance. The PMWB mechanism operates in two stages: wear prediction and dynamic write redistribution. First, the system monitors the memory’s wear status in real time, tracking how often specific cells are written to and assessing their wear-out progression. This data is used to predict future wear patterns and inform the system where redistribution is needed. Secondly, the system redistributes write traffic away from heavily worn cells ultimately, extending the overall lifespan of the memory. PMWB leverages adaptive machine learning models, particularly reinforcement learning (RL), to monitor and predict memory wear-out patterns. Reinforcement learning is ideal for this problem because it allows the system to learn optimal strategies for balancing memory wear out over time. In PMWB, the memory system is treated as an environment, and the RL agent interacts with it by taking actions (like redistributing write operations or adjusting approximation levels) to maximize a reward. In this scenario, the reward is a prolonged memory lifespan, and the RL agent is trained to optimize this goal by balancing wear across memory cells. The RL algorithm learns from experience by observing the wear patterns and adjusting its strategy dynamically. As the memory system undergoes more writes, the RL model refines its predictions of the cells likely to wear out first. Based on this predictive modeling, it redistributes write operations to less worn cells, thus balancing the wear across the entire memory. This delays the onset of errors and also ensures a more consistent performance over the memory’s lifetime. PMWB’s integration of approximation tuning is another innovative aspect. For applications where perfect accuracy is not always required such as multimedia processing, this trade-off is beneficial. The adaptive nature of PMWB’s reinforcement learning model means that it continuously improves over time, refining its predictions and redistributions as it encounters new write patterns. This adaptability is essential in memory systems where usage patterns may fluctuate due to workload. In fact, PMWB offers a robust and intelligent solution to the problem of uneven wear in non-volatile memory systems. By combining reinforcement learning-based wear prediction with dynamic approximation tuning, it extends memory endurance while preserving system performance. This novel approach represents a significant advancement in memory wear management, particularly for systems leveraging approximate computing, and has the potential to enhance the longevity and reliability of future NVM technologies.

Authored by
Dr. Marjan Asadinia (California State University, Northridge), Dr. Sherrene Bogle (California Polytechnic State University Humboldt ), and Rowena Quinn (Affiliation unknown)