Articles

Annual Forum Features Cutting-Edge Business Analytics Experts from Across the Nation

By Keith Morelli

Forum speakers

TAMPA (May 20, 2021) -- Artificial intelligence identifies, screens and in some cases treats children with dyslexia. Algorithms now detect high-risk patients to funnel them into preventative care programs. Other algorithms can predict a patient’s mortality based on random scans of chest X-rays.

The gist of the topics covered at the Florida Business Analytics Forum on Thursday was that artificial intelligence is making huge strides in the medical industry, and the future looks even brighter. But there is a long way to go to perfect the technology that may be relatively unknown among the general population whose only glimpse of artificial intelligence comes from “The Terminator” movies.

The forum, a popular event that was cancelled last year because of the COVID-19 crisis, returned today, drawing internationally renowned experts in artificial intelligence and ways of eliminating racial biases that may unconsciously creep into machine learning. More than 700 registered to attend the virtual event.

The forum this year focused on the future of analytics and artificial intelligence in medicine, said Balaji Padmanabhan, director of the Muma College of Business Center for Analytics and Creativity, which hosts the event each year. For the center, the forum is a signature event designed to put a spotlight on the movement of new ideas and knowledge into the industry by bringing in thought leaders who introduce cutting-edge ideas into the analytics space.

Among the speakers was Ziad Obermeyer, who has conducted nationally recognized research in the area of algorithmic racial bias in the health care industry.

“Everyone to some extent is worried about algorithmic bias,” he said. Identifying bias and managing it is the focus of his work. “The examples here are health care oriented but the lessons are general.”

Obermeyer, who teaches at the School of Health at the University of California Berkeley, researches the intersection of machine learning, medicine and health policy. His artificial intelligence research looks at the small number of patients who are chronically ill, patients who may slip through the cracks and end up in the emergency room, which ends up costing the industry more in the end.

Algorithms already are able to identify patients as high or low risk and programs are in place to offer preventative care.

“The need,” he said “is to target people who need care the most.” To do that, algorithms assign a score to all patients who receive primary care. However, they do not take into account people who don’t have health care or don’t follow up on primary care. That tends to be minorities or those in lower socioeconomic situations.

“The Idea is to get people who need extra help the extra help they need,” he said. “But there are some racial biases in some algorithms. Nevertheless, once you’ve detected those biases, you have a roadmap to fixing it. His work with companies that developed such algorithms has resulted in reductions in biases, he said.

“We’ve found these biases everywhere,” he said, “but we also found the solution as well.”

Hugo Aerts, one of the leaders in the field of artificial intelligence and medicine, spoke about deep learning in medicine and how artificial intelligence can be used to save lives and predict how long people live. Aerts, who teaches at Harvard Medical School, is the director of the Program for Artificial Intelligence in Medicine at Brigham and Women's Hospital.

His research focuses on the development and application of novel artificial intelligence approaches – deep learning in particular – for personalized medicine by integrating and analyzing medical imaging, pathology and genomic data.

He told forum participants that artificial intelligence is not a new field; that breakthroughs happened about a decade ago when computers began to match and in some cases surpass human performance.

“In the future, artificial intelligence will be involved in much more complex tasks such as surgeries, maybe even writing best sellers,” Aerts said. “We do expect major breakthroughs in medicine especially the fields of imaging fields in the coming years.”

One example of his research looked at predicting the future.

“Individuals age at different rates,” he said. His algorithm looked at tens of thousands of chest X-rays, to inspect all variables to predict which patients would become high risks. The algorithm accurately identified the ages of the patients and was able to predict future issues of potentially life-threatening problems.

“Of those identified as low risk, 96 percent were still alive after 12 years,” Aerts said, “Those identified as high risk, only 45 percent were living.”

Luz Rello, who suffers from dyslexia, has done innovative research to detect dyslexia early in children by using machine learning. She is the founder of Change Dyslexia, a social organization to screen and treat dyslexia, and an assistant professor in the Department of Information Systems and Technology at IE Business School at IE University in Spain.

She used eye-tracking data to first identify students with dyslexia. She found that students reading text reacted differently when spotting an error. She mounted research that involved 400 schools in Spain to identify and treat children with dyslexia.

“Dyslexia is a learning disorder,” she said at the forum. “Errors (in text), it turns out, are valuable research data. If there are errors in the text, people with dyslexia see errors and react differently than people without dyslexia.”

Identifying children with dyslexia is a major step, as often those children were left behind, being blamed for poor reading skills or being the products of inadequate education, she said. She has developed a course of treatment that she is introducing to schools and parents to help children overcome the disorder.”

The forum included talks, question and answer periods and power points to emphasize key points.

“This is not a typical conference,” Padmanabhan told the participants at the outset of the forum. “The audience here today is as much of an expert as us and we are looking forward to getting input from you as well.” 

To view a recording of the forum, click here.