Newswise – TROY, NY – A number of vulnerabilities, known collectively as deep learning adversaries, are preventing artificial intelligence (AI) from its full potential in applications such as quality improvement medical imaging and computer-assisted diagnosis. With the support of a National Science Foundation Teacher Career Development Program (CAREER) Award, Pingkun Yan, assistant professor of biomedical engineering at Rensselaer Polytechnic Institute, will lead a team of researchers in the development of new AI techniques that protect algorithms from these vulnerabilities, including contaminated data, malicious attacks or independent algorithms that interfere the ones with the others.
“We see great potential in AI, machine learning and deep learning in biomedical imaging,” said Yan, member of the Center for Biotechnology and Interdisciplinary Studies (CBIS) in Rensselaer. “We just need to build a system step by step to make it more robust, usable and understandable. ”
AI techniques, like the ones Yan and his team have previously developed and tested, have the potential to advance image reconstruction, image quality, computer-aided diagnosis and image-guided surgery. But opponents of deep learning, Yan said, remain an obstacle to widespread implementation. They could cause inaccurate or confusing image results.
“The adversary could cause the system to generate unwanted or unwanted outputs,” Yan said. “Clinicians can be confused by the outcome of the system, or a misdiagnosis can simply slip through. This could lead to significant medical errors in the diagnostic process and result in significant costs to our healthcare system. ”
While there are currently tools to protect algorithms from adversaries, Yan said they often come at the cost of reduced performance. The goal of this research, funded by a grant of nearly $ 550,000, is to solve this problem by developing AI techniques that are robust enough to guard against deep learning problems without interfering with the tasks that d other algorithms are trying to perform. The tools developed by Yan and his team will be able to detect contradictory images, correct these discrepancies and improve the quality of the information produced by the AI system. New algorithms developed by the team must take into account the entire system, not just the individual components, Yan said, so that accuracy is not sacrificed as robustness increases.
“This CAREER award is recognition of Professor Yan’s innovative research with broad applications in multiple fields,” said Deepak Vashishth, Director of CBIS. “For example, he and his team continue to advance the potential of AI to improve human health, while addressing persistent challenges that must be overcome along the way.
“We believe that such systems will help to make medical imaging more robust, more precise and will also boost the confidence of users, especially healthcare professionals,” said Yan.
About the Rensselaer Polytechnic Institute
Founded in 1824, the Rensselaer Polytechnic Institute is the first technological research university in the United States. Rensselaer encompasses five schools, 32 research centers, over 145 academic programs and a vibrant community of over 7,600 students and over 100,000 living alumni. Rensselaer’s faculty and alumni include more than 145 members of the National Academy, six members of the National Inventors Hall of Fame, six winners of the National Medal of Technology, five winners of the National Medal of Science and one award winner. Nobel Prize in Physics. With nearly 200 years of experience advancing scientific and technological knowledge, Rensselaer remains focused on solving global challenges with a spirit of ingenuity and collaboration. For more information, please visit www.rpi.edu.