A Campus Visionary
After nearly four years leading an advanced artificial intelligence initiative for the US Department of Defense, University of Massachusetts Amherst Professor Hava Siegelmann is thrilled to return to teaching this semester. “It’s very exciting; it feels like being a new professor,” she says. “I love teaching and I love students. That was the part that was missing while I was in Washington. It’s always so much fun to see their passion and watch them mature.”
Siegelmann, a professor of computer science as well as a core member of the neuroscience and behavior program, was on loan to DARPA (the Defense Advanced Research Projects Agency). Her most recent research, partly funded by DARPA and published in Nature in September, draws inspiration from the human brain to develop a new method to keep memories from fading in artificial intelligence neural networks.
When Siegelmann completed her service, she was awarded the Meritorious Public Service Medal—one of the highest honors the agency can bestow on a private citizen. The rarely awarded medal sits in a packing box in her basement as she works to transfer her undergraduate course on algorithms to Zoom and resumes leadership of a campus research lab.
A professor at UMass Amherst since 2001, Siegelmann came to her interest in learning and memory as a teenager in Israel. “From age 15 or 16, I was reading about how people learn,” she says. “I was curious about how memory is organized and how we can learn more things and not forget what we learned previously. I wanted to know how we can improve memory.”
Siegelmann first applied and was accepted to medical school, with the goal of becoming a psychiatrist. But she was also fascinated by mathematics and research, and so opted to study at Israel’s Technion, where she discovered that computer science could encompass her many interests. “For me, it was all about neural networks and machine learning even before we had names for them,” she says. She earned her master’s degree in computer science at Hebrew University and her PhD at Rutgers University.
Siegelmann has been lauded for visionary thinking throughout her career. “In order to have totally new ideas, you have to look at things in a different way,” she says. “There are so many things in computer science that we miss if we don’t think of it together with neuroscience. The basis of computer science is founded on wondering about how humans do things. Alan Turing and others who pioneered computer science were always thinking about the human brain. That approach was lost, but now computation combined with neuroscience is bringing that idea back.”
At UMass Amherst, Siegelmann heads the BINDS (Biologically Inspired Neural and Dynamical Systems) lab. The lab researches next generation AI and how to apply biologically inspired computing and computational methods to biology and medicine. Its research adds to our understanding of diseases including cancer, neurological diseases, and immunological diseases. “I want to do research for the good of humanity,” says Siegelmann.
She believes computer science and neuroscience can both be summed up in one word—flexibility. The limitation of today’s artificial intelligence, she says, is that computers cannot adapt to changing circumstances. The future of computer technology lies in computers becoming more flexible—more aware and more sensitive to circumstances. “If we take these traits into our lives, we can all be better learners, better teachers, and better researchers,” she says.
After her award-winning stint at DARPA, COVID-19 is freeing Siegelmann from the demands of presenting her research at international conferences, giving her more time for teaching and research. Back in Amherst, she says: “I just love my job. I’m lucky there is such a job that combines learning and memory, research and practice, and above all—students. The university is my place."