By Dave DeFusco
Artificial intelligence often feels like a distant or abstract concept鈥攕omething happening inside giant tech companies or futuristic labs鈥攂ut at the Katz School of Science and Health, artificial intelligence is being shaped into tools that address very real, very human challenges. Over the past year, faculty members and students across the school have worked on projects that bring clarity to medical decisions, protect people from digital threats, improve access to education and make complex systems easier to understand and use.
Rather than focusing on machines that 鈥渢hink like humans,鈥 Katz School researchers are concentrating on something more practical: using computers to recognize patterns, learn from experience and support better decisions. In simple terms, artificial intelligence here is about making technology more helpful, fair and responsive to the world as it actually exists.
鈥淲e鈥檙e not building AI for its own sake,鈥 said Honggang Wang, chair of the Department of Graduate Computer Science and Engineering. 鈥淲e鈥檙e building systems that solve real problems鈥攑roblems in healthcare, security, education and daily life鈥攚here smarter technology can make a meaningful difference.鈥
Bringing Advanced Healthcare Tools to Everyday Devices
Healthcare has been one of the most important areas of focus. One standout project is a mobile app called Pink Guardian, designed to help with early breast cancer detection. The idea is straightforward: allow people to upload mammogram images using a smartphone and receive instant feedback powered by artificial intelligence.
The app works by comparing an image to thousands of examples it has studied before, learning what healthy tissue usually looks like and how it differs from tissue that may need further medical attention. Along with a prediction, the app provides a confidence score, giving users a sense of how certain the system is. The technology runs directly on a phone, meaning it can be used in places without access to large hospitals or advanced computing systems.
鈥淭his kind of work shows how AI can bring medical expertise closer to patients,鈥 said Wang. 鈥淲hen you can run advanced analysis on a mobile device, you remove many of the barriers that prevent people from getting timely care.鈥
This project reflects a larger goal seen across Katz School research: bringing powerful medical insights closer to patients and clinicians, rather than keeping them locked inside specialized institutions.
Another healthcare innovation focuses on colorectal cancer. Detecting precancerous polyps during colonoscopies is difficult, and even experienced clinicians can miss subtle signs. Katz School researchers developed an image-analysis system that helps highlight potential polyps more clearly, improving the chances of early detection and treatment.
Additional projects explore lung cancer detection from CT scans, dietary monitoring using smartwatches and even 鈥渄igital twin鈥 health systems powered by smart rings. These tools aim to quietly collect useful health information in the background, helping doctors and patients respond to problems sooner rather than later.
鈥淚n healthcare, speed and accuracy matter,鈥 said Wang. 鈥淚f AI can help clinicians see patterns earlier or make sense of complex data faster, that can translate directly into better outcomes for patients.鈥
Protecting Sensitive Data in a Connected World
As healthcare increasingly relies on smartphones, wearables and connected devices, protecting sensitive information becomes just as important as analyzing it. Several Katz School projects tackle this challenge head-on.
One major effort focuses on cybersecurity education for mobile health technologies. New research-backed curricula are being developed to teach students how to secure wearable devices, protect wireless communications and ensure that artificial intelligence systems handle health data responsibly.
鈥淥ur goal is to train professionals who understand both sides of the equation,鈥 said Wang. 鈥淭hey need to know how these systems work, and they need to know how to protect them from misuse or attack.鈥
At the same time, researchers are developing technical solutions that allow medical data to be used without being exposed. One approach lets artificial intelligence models learn from information stored at different hospitals or research centers without ever moving the data to a single location. This makes collaboration possible while preserving patient privacy.
鈥淭hese methods are especially important in healthcare,鈥 said Wang. 鈥淭hey allow innovation to move forward while respecting ethical and legal boundaries around sensitive data.鈥
Fighting Financial Fraud with Smarter Systems
Healthcare is not the only area where security matters. As digital payments become the norm, financial fraud has grown more sophisticated. Traditional fraud detection systems often rely on fixed rules, which criminals quickly learn to bypass.
A team of Katz School cybersecurity students explored a more adaptive solution by combining two modern technologies: machine learning and blockchain. Machine learning allows a system to study past transactions and recognize patterns that may indicate fraud, even as those patterns change over time. Blockchain adds a secure, tamper-resistant record of transactions, making it extremely difficult for criminals to alter data without detection.
This project reflects a broader theme across Katz School research: building systems that can adapt. Instead of static rules, these tools learn from experience, making them better suited to fast-changing digital environments.
Detecting the Unexpected Even on Small Devices
In a world filled with sensors, smartwatches and connected machines, massive amounts of data are generated every second. Sometimes, the most important information is not the average value, but the unusual one: the data point that doesn鈥檛 fit the pattern. These 鈥渙utliers鈥 can signal early warnings, such as a failing machine part or the first signs of a medical emergency.
Many existing tools for detecting these anomalies require heavy computing power. Katz School researchers have developed lightweight alternatives that can run efficiently on low-power devices.
鈥淣ot every device has access to a data center,鈥 said David Li, program director of the M.S. in Data Analytics and Visualization. 鈥淲e鈥檙e focused on making models that are efficient enough to work where the data is actually being generated鈥攐n wearables, sensors and embedded systems.鈥
These efficient methods are designed for real-world environments where resources are limited. By catching problems early, they can prevent much larger failures later.
Understanding Creativity and Limits in Artificial Intelligence
Artificial intelligence has gained attention for its creative abilities, especially through image-generation tools that can produce realistic pictures from simple text prompts. Katz School researchers have taken a closer look at how these systems actually generate content.
Their work examines whether such models truly create new information or simply rearrange patterns they have already seen. This research helps clarify what today鈥檚 AI systems can and cannot do, offering a more realistic understanding of machine creativity.
鈥淚t鈥檚 important not to overestimate these systems,鈥 said Li. 鈥淭hey are powerful tools, but they still depend heavily on the data they were trained on. Understanding those limits helps us use them more responsibly.鈥
Securing AI Systems Against Hidden Threats
As artificial intelligence systems become more powerful, they also become targets for new kinds of attacks. One particularly dangerous threat involves secretly poisoning the data used to train an AI model, causing it to fail under specific conditions.
Katz School researchers developed a defense strategy inspired by biological vaccination. Their approach allows contributors to shared AI systems to protect models against hidden attacks, even without direct access to the model鈥檚 internal code.
鈥淭his kind of research is critical as AI systems become more distributed,鈥 said Li. 鈥淲hen many people contribute data, security can鈥檛 be an afterthought; it has to be built in from the start.鈥
Expanding Access to Education Through Technology
Education has been another major focus, particularly for communities with limited resources. One student-led initiative is developing an offline, AI-powered learning platform designed for classrooms without reliable internet access.
鈥淎I doesn鈥檛 have to mean constant connectivity,鈥 said Li. 鈥淪ome of the most impactful solutions are the ones that work under real-world constraints.鈥
Other education-focused projects include tools that help patients find clinical trials, systems that reduce barriers for underserved communities and platforms that help educators evaluate literacy programs around the world.
Making Large AI Models Easier to Adapt
Many modern AI systems are enormous and difficult to update. Katz School students have explored smarter ways to adjust only the most important parts of these models, allowing them to learn new tasks more quickly and efficiently.
鈥淭his is about sustainability,鈥 said Li. 鈥淚f every update requires massive computing resources, that limits who can use and benefit from AI.鈥
A Shared Commitment to Practical Innovation
Across all these projects, a common theme emerges. Artificial intelligence at the Katz School is not treated as magic or mystery. It's considered a practical set of tools鈥攑owerful, but grounded in real needs and real constraints.
鈥淲hat unites this work is purpose,鈥 said Wang. 鈥淲hether it鈥檚 healthcare, cybersecurity or education, the focus is always on how technology can serve people better.鈥
Li echoed that sentiment, pointing to the wide range of student-led projects. 鈥淭hese students aren鈥檛 just learning algorithms,鈥 he said. 鈥淭hey鈥檙e learning how to ask the right questions about the world and use data to answer them in meaningful ways.鈥
