It was at Drexel that alumnus Lex Fridman—now a research scientist at MIT working on autonomous vehicles—cultivated the yin and yang of engineering.
For those unfamiliar with the adroit use of third-century BCE philosophies at the College of Engineering, Fridman has a ready explanation. One professor fostered his “childlike” curiosity and joy in tackling big research questions. Another fostered the pragmatic and the practical side of research. The dual approach was so effective that Fridman never felt the need to look elsewhere, and fulfilled all three of his degrees at Drexel.
This Thursday, Oct. 4, Fridman puts it all on display at an open, one-hour seminar on “Human-Centered Autonomous Vehicles,” from 4 to 5 p.m. at Bossone 302. The talk is sponsored by the Department of Electrical and Computer Engineering. All are welcome.
Fridman’s lecture will feature a discussion on artificial intelligence (AI) and driver-state sensing, voice-based transfer of control, the annotation of large-scale naturalistic driving data, and the challenges of building and testing a human-centered autonomous vehicle.
“Deep learning is a way to teach machines to find patterns in data so that they can solve problems,” said Fridman during a phone conversation from his Cambridge, MA office. “In cars, that means collecting huge amounts of external data of driving scenes and making the AI system find patterns in those scenes in order for it to be able to drive, to avoid pedestrians, to steer. In other words, to make the car see.
“My particular work is in building cars that do this, what we call human-centered autonomous vehicles – cars that don’t just look out into the world but that have sensors looking at the humans inside the car to interpret what’s going on with them,” Fridman added. “The eyes are the windows to the soul, right? These sensors can detect where the driver is looking. They sense the cognitive-load through eye movement. They sense how preoccupied your mind is. That’s one of several things we’re interested in.”
Self-driving cars are one of the most obvious ways AI is starting to impact public life. What is not so obvious is how AI does it. That’s where machine learning comes in, along with the engineers behind it who seek to reproduce in machines what humans do exceptionally well – learn. From Fridman’s point of view, a car infused with AI is a potential partner in driving, maximizing safety by minimizing human flaws and its own flaws through a constant exchange of information.
In short, Fridman sees his task as no less than building a successful relationship between the human machine and the actual machine.
A Line of Fridmans
Fridman’s Drexel lineage is more pronounced than most. His father, Dr. Alexander Fridman, is the John A. Nyheim Chair Professor in Mechanical Engineering and Mechanics at CoE (MEM), as also the director of the Drexel Plasma Institute. His brother, Dr. Gregory Fridman, is an assistant research professor in Drexel’s School of Biomedical Engineering, Science and Health Systems. Drexel was, and is, a family affair.
Fridman earned his BS/MS in computer science in 2010, back when the degree program was still a part of the College. He earned his PhD in electrical and computer engineering in 2014. The two professors who had the most impact on his research path were Steve Weber, professor and ECE department head; and Moshe Kam, former ECE department head and now dean of the Newark College of Engineering at the New Jersey Institute of Technology.
In particular, Fridman commends Weber’s “childlike curiosity and passion for research” as providing the essential spark of creativity that so many engineering schools fail to value. Kam, alternatively, was the pragmatist; his credo to move through a problem incrementally inspired a yeoman’s diligence when tackling research questions.
“Lex Fridman, whom I had the pleasure to advise during part of his studies at Drexel, has a multi-tracked mind, overflowing with hypotheses, postulates, observations and theories,” said Kam of his former student. “For years my role as advisor has been to try to stem Alex’s deluge of arguments, concepts, beliefs and precepts. The goal was to have a manageable fraction of it formulated, analyzed, exemplified and presented to the technical community. The abundant remainder was to be properly documented and preserved in its original freshness for use in the next round.
“As is always the case with brilliant students,” Kam added, “I delighted to see how in time my role keeps diminishing, the advisor is relegated to the periphery, and soon enough I am learning from Lex much more than I teach him.”
One of the many problems Fridman worked on at CoE was a DARPA-funded project on active authentication, or, using AI to determine the identity of an individual based on behavioral patterns. It was a problem many were working on at the time, Fridman said. Companies wanted to bypass the need for passwords and other manual IDs. How an individual uses a phone, dials a number, or manipulates a keypad along with facial markers – all of it adds up to an ID system that would obviate the need for passwords and provide additional safeguards.
The research formed the basis of his PhD thesis.
A Word of Advice
After Drexel, Fridman spent six months at Google continuing the authentication work he had started as a PhD candidate. He found the work to be pleasant and affirming, if not the challenging environment he expected it to be. Ultimately, it was academic research that held him in thrall. He was offered a position as a research scientist at the Massachusetts Institute of Technology, which he took without looking back.
“I like the chaos of research and the academic environment,” Fridman said of his decision to leave Google. “My values are really about doing the best, coming up with the coolest, most interesting ideas you can. For me, the academic environment works best, where friends and colleagues around me can kind of create that kind of atmosphere.”
Today, he sees his years at Drexel as among the most rewarding of his life. Asked if he has any advice to pass on to undergraduates, Fridman responded generously, urging students to take advantage of their college years to pursue intellectual questions they might not later have the time for.
“Don’t see your undergraduate work as a stepping stone to a career, or even the next step,” said Fridman. “See it as one of the best times in your life to grow, to learn, to ask fascinating questions, to explore the mysteries of the universe. If you see it that way, the place you end up is going to be the place you want to be.”