Even The Sky is Not the Limit with AI and Data Science: Rama Chellappa Faculty Spotlight

Since the Industrial Revolution, maximizing the usefulness of new technologies has meant continually adapting and refining them, and this will also be true for AI and its applications,” — Rama Chellappa, Johns Hopkins Bloomberg Distinguished Professor of electrical and computer and biomedical engineering, Interim co-director of the Data Science and Artificial Intelligence Institute.

Rama Chellappa originally aspired to be a teacher, like his father, but that changed when he was 15 years old, while living in Chennai, and listened to a radio broadcast of the Apollo moon landing. Impressed by this incredible feat of engineering and science, he decided he wanted to be an engineer.

This led him to studying at the University of Madras and the Indian Institute of Science before pursuing a PhD in the United States at Purdue University. He began his teaching career at the University of Southern California and then spent 29 years on the faculty at the University of Maryland before joining Johns Hopkins University in 2020 to pursue research combining the fields of AI, machine learning, engineering, and medicine.

“My parents were very nervous about my desire to join the teaching profession, but I was very determined to work in academia,” Chellappa says, and he is grateful for the opportunities his career continues to afford him. “I really enjoy my work and my students. It’s just an amazing job and Johns Hopkins is a great place to work and collaborate.”

Since the early 1980’s, a major focus of Chellappa’s AI-related research has been in the fields of pattern recognition and computer vision. Pattern recognition has many applications, from fingerprint to face and speech recognition. Computer vision, on the other hand, is processing images and videos with the goal of understanding a scene, the objects, and their interactions.

Chellappa was introduced to the field of pattern recognition and artificial intelligence in the late 1970s, and his doctoral dissertation was on small generative AI. At the time, he says the computing facilities he used could generate only 32×32 images. “Over the years, pattern recognition has morphed into machine learning and data science,” Chellappa explains, noting that AI, data science, and machine learning are related disciplines that all rely on the possibilities presented by large amounts of data.

He is also interested in robotics, a discipline in which AI, including computer vision and sensors, can be applied to the practice of medicine.

One of his earliest memories of AI applications to medicine was in 1973, when a solution to a blood infusion infection problem was never deployed because of the brittleness of the rules-based approaches that were popular then, as well as concerns about privacy.

Chellappa’s interests in AI’s applications to health and medicine is enabled thanks to the enormous amount of medical data that exists and can be analyzed. “Since 2012, AI has been largely data-driven, and the more data you have, the better you do with AI,” he says.

At Johns Hopkins, he collaborates with researchers in the Whiting School of Engineering and clinicians at Johns Hopkins School of Medicine on projects that are addressing big-picture issues in AI and machine learning. “People are concerned that AI needs to work everywhere and for everyone,” he says. “For example, if a machine learning program is developed in India, will it be equally effective in U.S.? We need to be able to detect and mitigate bias if AI is to be a robust technology that does good for all.”

Data Science and AI Institute

In August 2023, Johns Hopkins University announced the creation of the Data Science and AI Institute (DSAI), a university-wide effort focusing on the emerging applications, opportunities, and challenges presented by data science, machine learning, and AI. As the institute’s interim co-director, Chellappa works with people across university divisions who are pursuing high-impact, cross-disciplinary research. One aspect of the initiative is significant faculty growth, including appointing 30 new Bloomberg Distinguished Professorships, positions like Chellappa’s that have appointments in at least two university divisions.

Under the auspices of DSAI, Johns Hopkins has hosted high-profile events including congressional panels with U.S. senators, symposia, and a panel with renowned tech journalist Kara Swisher. Soon, they will join the Science Diplomacy Summit, where they will meet embassy representatives to discuss AI in their respective countries and how they can work together to make AI a force for good in diplomacy.

“Plans are underway to construct a building on Johns Hopkins Homewood campus, which will be a modern facility and base for the university’s AI research. JHU is also significantly growing its compute capabilities which will enable the collection and curation of trusted data sets, which will be made available to researchers,” Chellappa says. “DSAI will break the barriers between departments, institutes and centers. It is already making Hopkins the place for doing cutting-edge data science and AI research.”

The Future of AI

Chellappa likens AI to a two-year old child, albeit one that is very good at calculus.

“Let the child grow, and let’s not demand everything from it as it is still learning to walk and talk, he says.  While people may have concerns about AI, fueled by images like the Terminator, Chellappa is confident that AI will improve quality of life on a global scale. Many people don’t go to doctors routinely, only when they’re sick. AI will allow us to maintain a record of what patients are doing and provide diagnoses in remote areas where doctors may not be available.”

AI will enable us to monitor patient health. Through wearable tech, it can do things like check blood sugar levels and track heart health and can even provide disease diagnosis in remote areas where doctors may not be available.

“Twenty years ago, at a panel, somebody suddenly threw a question at me: what is your moon shot? My moon shot in education is one physical teacher and 30 virtual teachers so that a classroom of 30 students can have personalized instruction for every student. AI can follow what you do, what kind of lessons you need, and give encouragement,” Dr. Chellappa continued. Students are not using libraries much these days, so AI would be useful in knowledge transfer.”

 “Don’t think of AI as how a person can buy a Tesla and drive comfortably to work. Look at how an AI-driven car can help a person with vision or mobility issues travel to the fullest extent possible. AI is for helping and improving the lives of the vulnerable, the needy, and really all of us around the globe.”