The Northwestern Mutual Data Science Institute has awarded $675,000 to researchers at its university partners, UW-Milwaukee and Marquette University.

The awards are part of three innovative programs and engagement opportunities that fall under the institute’s Center of Excellence initiative.

Focused around five Center of Excellence areas – artificial intelligence (AI); AI, data bias and ethics; behavioral economics; financial literacy; and health and wealth inequities – each of these programs will serve to accelerate a robust ecosystem for research and innovation, talent pathways, and community outreach and partnerships. To foster broader inclusion and cross-institution collaboration, NMDSI considered nearly 40 proposals for cohort-modeled opportunities, internships, research engagement and curricula development.

NMDSI awarded up to $500,000 as part of the Paving ROADS Seed Fund Program, $100,000 on behalf of the Pioneer Collaborative Curricula Program and $75,000 for the NMDSI Student Research Scholars Program.

Award winners and their projects are:

Paving ROADS Seed Fund Program

This annual program aims to support new research partnerships to forge short- and long-term engagement among NMDSI-affiliated faculty and data science experts across disciplines while strengthening cross-campus research to use data science for social and societal impact in Milwaukee and beyond.

• Does More Money Buy Better Health? Machine learning-based novel causal estimation methods for analyzing evidence from 500 largest U.S. cities including Milwaukee
• Avik Chakrabarti, associate professor of economics at UWM, and Sabirat Rubya, assistant professor of computer science at Marquette

Abstract: A fundamental problem with making international comparisons of well-being is that there is no standard metric for health, as opposed to the very standardized measurement of income. In this project, standardized measures of health disparities and income inequalities obtained from granular data on census tracts will be used to identify if disparities in health and inequalities in income are prevalent.

• Improving Student Success using AI and Machine Learning
• Jun Zhang, professor of electrical engineering and computer science at UWM

Abstract: UWM is a public access university that provides great opportunities for a diverse population of students, especially first-generation students. However, retention and prompt graduation have been a great challenge. The purpose of this project is to develop machine learning algorithms and software that can provide automated early prediction of the students’ course outcomes and identify actions that can help improve performance and avoid failure.

• Towards a Comprehensive Framework and Repository for Documenting and Assessing Biases in Using Large Language Models in Health
• Lu He, professor of public health (UWM), Priya Deshpande, assistant professor of electrical and computer engineering (Marquette), and Praveen Madiraju, assistant professor of computer science (Marquette)

Abstract: The recent emergence of large language models such as ChatGPT brought increasing excitement in using AI in the health domain. However, there is evidence that LLMs may have real-world biases such as gender, racial and age biases that could harm patients, especially those who are underrepresented. The goals of this project are to generate a comprehensive framework that builds on how to conceptualize and assess biases in using LLMs in the health domain and create and publicize a repository that documents and continuously updates these biases.

• Speech-Mediated Robot-Assisted Therapy: Upper limb rehabilitation using smart assistant empowered by OpenAI
• Mohammad Rahman, associate professor of mechanical engineering (UWM), Susan McRoy, chair and professor of computer science (UWM), Inga Wang, associate professor of occupational therapy (UWM), and Sheikh Iqbal Ahamed, professor of computer science (Marquette)

Abstract: Aimed at revolutionizing robotic rehabilitation therapy, the project’s core objective is to develop an intelligent user interface that harnesses the contextual awareness of OpenAI and voice-activated technology, exemplified by Amazon’s Alexa. This innovative approach seeks to overcome the limitations of current robot-assisted rehabilitation therapy, including restricted customization, complexity, diminished human interaction, constrained motion and task variability. By promoting patient-centered care and leveraging technological advancements, the project has the potential to transform rehabilitation therapy delivery and enhance the quality of life for patients.

• Federated Learning for Educational Research
• Tian Zhao, associate professor of computer science (UWM)

Abstract: The student data used for training machine-learning models is privacy sensitive, which must be protected against malicious attack and accidental misuse. This makes it difficult to collect large amounts of student data, which is necessary for training effective machine-learning models. In this project, we will develop a distributed machine-learning system for educational research, where student data used for training machine-learning models is not transmitted to a central server. This would ensure that the student data is never at risk of exposure or misuse due.

• Multi-Modal Ranking Models for T-Cell Receptor-Epitope Binding Prediction
• Keke Chen, associate professor of computer science (Marquette)

Abstract: Accurate identification of TCR-epitope recognition is critical for understanding a functioning adaptive immune system and designing immunotherapy methods. Since the traditional web-lab approach is slow and expensive, machine-learning-based TCR-epitope binding prediction has become an emerging approach complementing the traditional one. This research will address the scarce training data issue and help develop several methods to explain the model prediction and link the model output to critical amino acids and structural features.

• UbiWhite: Smartphone-based non-invasive white blood cell counting system from fingertip videos
• Sheikh Iqbal Ahamed, professor of computer science (Marquette)

Abstract: This project introduces a groundbreaking solution aiming to transform health care by presenting a noninvasive, smartphone-based system for real-time counting of white blood cells. White blood cells, critical components of the immune system, defend the body against infections and foreign substances. Our proposed system circumvents limitations such as invasive blood draws and laboratory testing by leveraging the optical and magnetic properties of white blood cells to enable real-time counting without the need for blood samples. Ultimately, this system will enhance patient outcomes, facilitate early disease detection and contribute to the advancement of personalized and precision medicine.

Pioneer Collaborative Curricula Program

The Pioneer Collaborative Curricula Program seeks to introduce and embed emerging areas of data science into the curricula of our two university partners. We will engage NMDSI faculty to support the development of novel and advanced data science courses to supplement or expand existing curricula.

• Data Science for Public Policy
• Amanda Heideman, teaching assistant professor of political science (Marquette)

Abstract: The course “Data Science for Public Policy” aspires to address the increasing demand for novel interdisciplinary learning opportunities that illustrate how the latest developments in data science can help to improve the analysis of public policy. This 15-week course, open to graduate and undergraduate students alike, will help participants develop vital, hands-on experience with analyzing real microdata and applying data science skills to analyze real social problems, using large quantities of data from a variety of sources.

• Digital Health Using Data Science
• Sheikh Iqbal Ahamed, professor of computer science (Marquette)

Abstract: Sheikh Iqbal Ahamed’s “Introduction to Digital Health Data Science” class strives to offer a holistic understanding of the intricate intersection between health care and data science. In our rapidly evolving health care landscape, the strategic utilization of digital health data stands as a linchpin for well-informed decision-making, nuanced policy formulation and the enhancement of patient outcomes. The course accentuates the escalating prevalence of chronic diseases, the challenges posed by an aging population and the overarching strain on global health care systems. The goals of this course are to empower health care professionals to adeptly employ data-driven approaches for personalized patient care, provide researchers with the tools to extract meaningful patterns from expansive datasets and equip policymakers to make informed decisions grounded in statistical evidence.

• “Speaking Data” Course Sequence
• Erin K. Ruppel, associate professor of communication (UWM)

Abstract: The Speaking Data Course Sequence is crafted to provide students with a systematic and robust comprehension of the art of speaking data, enabling students to master the nuances of data interpretation, navigate all stages of transforming raw data into actionable insights and adeptly communicating these insights to diverse audience and stakeholders. Particularly, the sequence aims to equip students with up-to-date knowledge of effective communication in data science and rich hands-on experiences to become proficient and ethical data communicators in contemporary society and the workplace, where AI, machine learning and other emerging technologies are fundamentally redefining traditional job roles.

• Business Seminar: Data Analytics, AI, and Innovation
• Joan Shapiro Beigh, teaching professor of organizations and strategic management (UWM)

Abstract: This seminar will focus on a practical, hands-on study of the topics of data analytics, artificial intelligence, innovation and creativity. Students will perform critical analyses of business problems and organizations, as well as complex problems such as pressing issues related to pro-social, environmental, and business-related issues that matter to students today. The creativity and innovation components of the course will serve as the catalyst to engage students from the liberal arts and other creative, non-STEM, disciplines; it will enable them to feel confident in their ability to succeed in the class and to work collaboratively with students who have more quantitative backgrounds.

Student Research Scholars Program

The NMDSI Student Research Scholars Program seeks to engage students from partner institutions in data science research, working with NMDSI-affiliated faculty and data science experts in and cross disciplines to give them a hands-on experience in the application of data science. Some of the projects will include working with faculty on sponsored research and some will be in the nonprofit community of Milwaukee under the supervision of a faculty mentor. The program will provide students with a stipend over the semester.

• Predicting Risk of Obstructive Sleep Apnea and Cancer Susceptibility Post Long COVID Using Integrative Data Analytics and Artificial Neural Network Insights
• Manoj Rajesh Purohit (Marquette)

Abstract: Long COVID, characterized by persistent symptoms following COVID-19, presents notable health concerns. Its intersection with obstructive sleep apnea is significant due to shared issues such as irregular breathing, disrupted sleep and cardiovascular risks. This research aims to investigate the increased risk of obstructive sleep apnea in long COVID patients, exploring connections between long COVID severity and the development of obstructive sleep apnea influenced by chronic diseases and disruptions in sleep patterns. The primary objective is to examine obstructive sleep apnea risks in individuals with long COVID compared to those without.

• Crafting Digital Story Narratives Using GenAI to Facilitate Social Support for Postpartum Depression
• Farhat Tasnim Progga (Marquette)

Abstract: Social support is crucial for managing postpartum depression symptoms. Women without adequate social support are five times more likely to experience postpartum depression. This project will explore the effectiveness of digital storytelling in the context of online social support for women with perinatal mental health illnesses through a series of empirical and design inquiries.

• Rating of the Importance of Personal Information in Cybersecurity
• Tianyang Xiong (Marquette)

Abstract: This project revolves around assigning varying levels of vigilance to different types of personal information in the context of personal cybersecurity. The proposed approach involves analyzing data related to cybercrimes targeting individuals to identify prime targets for fraud, the types of information fraudsters seek, data collected by perpetrators to obtain other personal details and the success rates of cybercrimes targeting specific information. This analysis aims to establish a rating system for the importance of personal information, facilitating targeted and efficient preventive measures.

• How AI Technologies are Reshaping and Challenging Traditional Business Practices in Milwaukee
• Max Hartounian (UWM)

Abstract: The primary goal of this program is to help students understand how AI technologies like ChatGPT are reshaping and challenging traditional practices across multiple sectors. Students will develop a qualitative, experiential component for this project where they interview leaders at Northwestern Mutual and other Milwaukee companies to obtain real-world perspectives.

• Vision-Driven Robotic Assistance for Enhanced Independence in Daily Activities
• Md Tanzil Shahria (UWM)

Abstract: In the United States, where disabilities impact an estimated 61 million adults, the need for sophisticated assistive technologies is increasing. This research will address this critical need by designing a vision-based control system for a six degrees of freedom robotic manipulator tailored to enhance activities of daily living for those with mobility challenges. The goal of the research is to set a new benchmark in aligning technological innovation with real-world human needs and hold the potential to be adapted to other research fields for advanced robotic control.

• The Question Machine
• Anirudha Subrata Mitra (UWM)

Abstract: Scientific thinking often solves an existing question and must be predictable. The “problem of the problem” is that asking creative questions is the linchpin of the quality of research across the sciences, just as the best of art “does things” that make us move and feel moved; yet we posit that it is useful to consider that what each teaches and celebrates typically tends more toward either utility or novelty as an entry point. The goal of this project is to create an AI-generated model built to ask creative questions and generate multimodal outputs – both finding at least one interesting and useful question to ask and show some of the results as poetic works of art.

• Social Factors and Patient Perceptions Preventing Them from Accessing Health Care by Race and Ethnicity
• Raisa Nkweteyim (UWM)

Abstract: Background U.S. Census data revealed that in 2022, there were 7.9% of people who did not have health insurance living in the U.S. The Affordable Health Act made health insurance more affordable for households with income within the federal poverty level, but it is not well understood what other social factors and personal perceptions exist that hinder people from accessing health care. This study aims to understand what patient characteristics are associated with the unwillingness to access healthcare.

• Ethical AI: Challenges and Opportunities
• Juliana Hirt (UWM)

Abstract: Applications of artificial intelligence have become ubiquitous in daily lives. While many embraced AI with excitement, ethical concerns of AI have gained increasing attention. Ethical AI refers to responsible AI use that complies with established ethical principles. The goal of ethical AI is to optimize the strengths of AI while reducing risks and preventing harm. The outcome of this project will be an article identifying the challenges and opportunities of implementing ethical AI in information services with an overall goal of informing users on how ethical AI may shape their experiences.