What if we may higher anticipate how a illness like Alzheimer’s evolves and information extra personalised care selections earlier than circumstances worsen? That query helps drive work at Temple University, the place Xinghua Mindy Shi, affiliate professor of pc and data sciences on the Institute for Genomics and Evolutionary Medicine within the College of Science and Technology, is tackling one in every of healthcare’s most troublesome challenges: predicting and managing dementia development over time. The bottleneck for her research? Computing energy.
That limitation is now starting to shift via Pennsylvania’s new Keystone AI + Quantum Factory, which Temple helped launch as a founding member. The statewide consortium brings collectively universities, business companions and public-sector leaders to develop entry to superior computing infrastructure wanted for large-scale AI and quantum research.
“This kind of computing infrastructure dramatically accelerates our research,” defined Shi. “Some of the large-scale AI training and simulations that might previously have taken us weeks can now potentially be scaled down to a couple of days or even several hours.”
Shi’s work is a part of broader research exercise underway throughout Temple, the place college and researchers are exploring how AI, quantum applied sciences and superior computing will be utilized to urgent societal challenges.
“The Keystone AI + Quantum Factory is creating the kind of shared infrastructure that no single institution could build on its own,” mentioned Josh Gladden, Temple’s vice chairman for research. “Research like Mindy’s demonstrates how access to advanced computing can help accelerate work on urgent challenges faced by society by supporting the scale of computing that today’s research demands.”
Advancing dementia care via expertise
Shi’s research on the Keystone Factory, “Trustworthy Agentic AI for Personalized Dementia Management,” focuses on a central problem in dementia and Alzheimer’s illness care: The illness develops step by step over a few years, usually producing incomplete, inconsistent and continually altering indicators that make it troublesome to anticipate how a affected person’s situation could evolve. Her work goals to develop an agentic AI system able to constantly modeling a affected person’s situation over time, accounting for uncertainty and serving to information extra personalised care selections.
With a background in pc science and machine studying, Shi has spent years making use of superior computing applied sciences to genomics and healthcare challenges. Dementia research, she defined, stood out due to each the complexity of the illness and the profound impact it has on sufferers and households over lengthy intervals of time, significantly as getting old populations proceed to drive rising charges of dementia worldwide.
“One of the biggest challenges with dementia and Alzheimer’s disease care is that patients and families often do not know what is coming next,” mentioned Shi. “Symptoms can develop gradually over many years, and disease progression can vary from person to person. What we are building is a system that can continuously integrate information over time and learn evolving patterns, helping to improve disease management and care.”
To develop and refine the system, Shi’s group is coaching AI fashions utilizing large-scale simulations and augmented datasets to replicate how dementia progresses throughout completely different affected person populations. Those efforts will enable her group to check interventions, adapt methods and refine fashions earlier than they are often utilized in real-world scientific settings. Shi emphasised that one vital a part of reliable AI is knowing uncertainty.
“In healthcare, it is not enough for a system to simply make predictions,” she defined. “Clinicians also need to understand how reliable those predictions are, when the system may be uncertain and where human oversight is still needed. That is a major focus of our work.”
While this research at present depends on superior AI fashions and large-scale computing infrastructure, Shi mentioned rising quantum applied sciences may finally assist course of and mannequin more and more complicated datasets at a good larger scale.
Building expertise and increasing functions
In addition to supporting cutting-edge research, the Keystone Factory can be designed to assist put together college students and researchers for fields the place AI, quantum applied sciences and superior computing have gotten more and more central. For Temple, that expertise pipeline growth mission is carefully tied to research itself.
Students engaged on tasks like Shi’s are benefiting from expanded entry to the sort of large-scale computing energy that’s more and more defining business. That added capability permits researchers and college students to run bigger simulations, course of extra complicated datasets and pursue tasks that may in any other case be constrained by computing limitations. “Initiatives like the Keystone AI + Quantum Factory help expand what we are able to do as computing demands continue to grow,” Shi added.
Pennsylvania Department of Community and Economic Development Deputy Secretary for Technology and Entrepreneurship Jen Gilburg mentioned establishments like Temple are central to the consortium’s long-term objectives round innovation, workforce growth and financial competitiveness. “Temple University and other research institutions across the commonwealth will play an important role in helping Pennsylvania compete in rapidly evolving AI and quantum fields,” Gilburg mentioned. “The Keystone AI + Quantum Factory is designed to turn research into new companies, high-quality jobs and economic opportunity across the commonwealth.”
At Temple, Gladden echoed these sentiments, emphasizing the significance of guaranteeing Temple researchers and college students stay aggressive in quickly evolving fields which might be reworking research and business.
“This is about positioning our researchers and our students for where the field is going,” Gladden mentioned. “The ability to work at this scale, with access to advanced computing and collaboration across institutions, is what will define the next generation of discovery and how quickly those discoveries translate into real-world impact.”