Concrete that would absorb carbon dioxide from the environment. A new technology to help quicken the pace of drug discovery, or to diagnose illnesses.
Ten faculty teams studying a range of ways to harness artificial intelligence and machine learning technology to improve our world will receive funding to start projects like these and many others this summer through the University of Miami’s Laboratory for Integrative Knowledge, or U-LINK.
Funded by the Office of the Vice Provost for Research and Scholarship, the U-LINK program offers interdisciplinary faculty teams up to $100,000 to spend a year working on a project that ideally will be shared through research journals and eventually benefit the general public.
Founded in 2018, the U-LINK program aims to foster collaboration and innovation between faculty members across the University’s three campuses by tackling real-world problems. This year’s funding theme centers on artificial intelligence and its applications to construction and health care innovation, among other areas. Past years have offered grants focused on climate change and social equity, to name just a few themes.
“The teams awarded provide potential solutions to challenges through creative interdisciplinary collaborations and approaches, including strategic engagement of academic, community, and external partnerships,” said Dr. Maria Alcaide, vice provost of research and scholarship. “These teams will advance our knowledge through the development of tools, methods, and strategic partnerships to further advance research with the use of AI as we continue to evolve as a leading innovative institution.”
Improving concrete’s durability and sustainability
Concrete. It is the most prevalent man-made material in the world, yet creating it causes 8 percent of the world’s carbon emissions today. It also cracks quite a bit, as a result of water or movement.
“Discovery of Catalysts to Accelerate Carbon Capture in Concrete”
While making concrete typically leads to carbon dioxide emissions, concrete also naturally absorbs some of this carbon dioxide in a process known as carbonation. The process is so slow that its benefits do not offset the carbon emissions it produces. This team is using artificial intelligence, combined with experiments and modeling, to speed up the process of carbonation and lower concrete carbon emissions. The researchers hope to transform concrete into an element that pulls harmful carbon dioxide from the atmosphere more efficiently, and to “transform our nation’s infrastructure from silent contributors to climate change into active and vast carbon sinks.”
Principal investigator: Rajeev Prabhakar, professor of chemistry. He is joined by Luis Pestana, assistant professor of civil and architectural engineering, and Prannoy Suraneni, assistant professor of civil and architectural engineering.
“AI agent to engineer virus-based self-healing concrete”
By harnessing the power of advanced generative AI, this team hopes to marry their knowledge from the study of biology—specifically viral capsids, or the protective shells of viruses—with concrete science to develop a revolutionary type of self-repairing concrete. This concrete would contain healing agents, which could be released when the concrete cracks. A new AI tool called CUNAO will play a pivotal role in helping the researchers blend these two sciences into a biologically inspired solution to extend the lifespan of buildings and infrastructure with minimal maintenance.
Principal investigators: Pestana, along with Antoni Luque Santolaria, associate professor of biology.
Health Care Innovations
“Deep Learning for Early Detection of Alzheimer’s Disease and Related Dementias”
With millions of baby boomers in their golden years, and many more to come, the health care industry is bracing for an influx of neurodegenerative diseases. To illustrate this, an estimated 40 million people across the world have Alzheimer’s disease and related dementias, with this number expected to triple by 2050. This team's goal is to develop a series of biomarkers for Alzheimer’s disease and related dementias, along with Parkinson’s disease, multiple sclerosis, diabetic neuropathy, and optic neuropathy. This first phase will be focused on the development of AI methods for early detection of Alzheimer’s and related dementias using high-tech medical images.
Principal investigator: Xiaodong Cai, professor of electrical and computer engineering. He is joined by Dr. Jianhua Wang, professor of ophthalmology at Bascom Palmer Eye Institute and co-director of its Experimental Imaging Laboratory; Liang Liang, assistant professor of computer science; and Dr. Hong Jiang, associate professor of ophthalmology and neurology at Bascom Palmer Eye Institute.
“Artificial Intelligence for Covert Volitional Eye-Response Test”
Up to 40 percent of patients in a coma with acute brain injuries have eye tracking in response to visual stimuli, indicating higher levels of consciousness. However, this tracking is often not detected by medical professionals at a bedside assessment. Many times, this leads family members to decide that their loved one should no longer be on life-sustaining therapies after a discussion with medical professionals. University researchers believe there is a better way to discern brain activity—through eye movement. This team will explore using computer scanning to more accurately identify eye tracking, indicating higher levels of consciousness and to better assess a patient’s prognosis of recovery after acute brain injury.
Principal investigator: Dr. Ayham Alkhachroum, assistant professor of neurology, physician in the division of neurocritical care at the Miller School of Medicine, and principal investigator of the Hacking Consciousness lab. He is joined by Mohamed Abdel-Mottaleb, professor and chair of electrical and computer engineering; Tulay Koru-Sengal, professor of biostatistics in the Department of Public Health Sciences at the Miller School; Brian Manolovitz, a doctoral student in computer science; Mostafa Abdel-Mottaleb, a doctoral student in biomedical engineering; Gabriela Aklepi, a student at the Miller School; and Pablo de la Fuente, a graduate student in economics and computer science.
“Stable Biopharmaceutical Formulation Design”
This team will use machine learning techniques to create predictive models for protein-based monoclonal antibody treatments, which are growing in popularity to treat patients with cancer, rheumatoid arthritis, autoimmune diseases, asthma, and other conditions. This could help dramatically speed up the formulation and drug development process in the biopharmaceutical industry.
Principal investigator: Samiul Amin, professor of practice in the chemical, environmental and materials engineering. He is joined by Yelena Yesha, professor of computer science and Knight Foundation Endowed chair of Data Science and Artificial Intelligence.
“Artificial Intelligence for the Diagnosis and Evaluation of Skin Cancer”
At a time when dermatologists are struggling to meet patients’ demand for their expertise, this team will evaluate artificial intelligence tools for the identification and classification of skin cancer. The models created in this effort can be applied to the diagnosis of other dermatologic diseases, such as hair and nail disorders, guided by the availability of large-scale datasets. They may also create scalable improvements that would be applicable throughout dermatology, and more broadly, to health care in general.
Principal investigator: Dr. Keyvan Nouri, professor of dermatology, ophthalmology, otolaryngology, and surgery at the Miller School of Medicine. He is joined by Yesha; Dr. John Tsatalis, resident at the Miller School in dermatology and cutaneous surgery; Abdel-Mottaleb; and Dr. Scott Elman, assistant professor of clinical dermatology and cutaneous surgery at the Miller School.
“Advancing Histopathological Diagnostics”
This team hopes to address a pivotal challenge in medical imaging: the need for high-quality, diverse datasets to accurately interpret and diagnose diseases. Traditional methods of creating computer models based on existing data are limited, which in turn, restricts the potential for models to transfer this information to real-world applications.
Principal investigator: Himanshu Arora, research associate professor of urology. Working with him is Cheng-Bang Chen, assistant professor of industrial and systems engineering.
“Using AI to predict real-time speech and disability in the diverse inclusive preschool classroom”
Affecting up to 8 percent of preschoolers, early language delays can have major long-term consequences on a child’s development and academic achievement. As a result, this team aims to pinpoint behaviors that could shed light on whether young children may have a language delay or developmental disabilities, such as autism spectrum disorder, in order to make sure these children receive critical interventions to support their development in school. To do this, they will use existing information about preschoolers’ behaviors collected in three preschools across Miami-Dade County and use AI to analyze classroom behaviors such as movement and vocalizations to predict language delays and disability status. This will help to identify children who may need added support to help them thrive in the future.
Principal investigators: Vanessa Aguiar-Pulido, assistant professor of computer science, and Lynn Perry, associate professor of psychology. They are working with Daniel Messinger, professor of psychology, and Chaoming Song, associate professor of physics.
Earth Sciences
“Improving coral larval recruitment using engineering, biophysics, and generative AI”
Artificial coral reefs built from man-made materials are becoming a popular way to restore reefs damaged by climate change, pollution, and ocean acidification. The long-term success of artificial reefs depends on their ability to recruit coral larvae, and to provide the best conditions for coral larvae to settle on and grow. This team will explore, using AI and experiments, how coral larvae interact with their local physical environment on artificial cementitious substrates in the lab. They will also investigate how this interaction determines the settlement preference and location of baby corals.
Principal investigator: Suraneni. He is collaborating with Vivek Prakash, assistant professor of physics; Erotokritos Skordilis, lecturer of business technology; and Jenna Efrein, senior lecturer of glass art.
“Humans and AI in concert (HAIC): Musicians' perceptions of acceptability and self-efficacy when cocreating with artificial intelligence”
Over the last five years, AI tools for music making have exploded. This team will explore how putting AI tools in the hands of music creators effect the participants and empower them to help shape the narrative of how AI can and will be used in the music industry.
Principal investigator: Tom Collins, associate professor of music engineering technology. He is working with Raina Murnak, assistant professor of modern artist development and entrepreneurship, and Christopher Bennett, associate professor of music engineering.
