Vladimir "Vova" Braverman has recently joined Rice University as the Victor E. Cameron Professor of Computer Science. At Rice, he will direct a research team in addition to teaching seminars and courses in his areas of expertise—algorithms and their applications for machine learning, computer networks, and medical imaging.
“I have been fascinated by algorithms from the moment I was introduced to them as an undergraduate,” he said. “The fact that through pure thinking—throwing something on a white board and working through the mathematical representation of a problem—you can improve a person’s life, this is exciting. In addition to designing theoretical algorithms during my academic career, I was also fortunate to work in industry and witness the impact of these algorithms across various fields.”
He enjoys teaching students to better interpret and translate actual problems into mathematical equations and then explore those equations from various angles as they search for algorithmic solutions. Although he thrives in the area of teaching and researching theoretical aspects of algorithmic problem solving, he is even more excited when his work is used to solve challenges faced by other people.
“In areas ranging from genomics to computer networks, we can formulate a problem in a mathematical way such that a real-life challenge becomes an abstract idea, graph or other object we can study. Then we can sit and stare at this abstract object on the white board, look at it from different angles, and "play" with the problem. Computer scientists who study algorithms actually get to play with their toys! But through this seemingly childish exercise comes serious solutions that can impact and improve our daily lives,” said Braverman.
“For example, one of my areas of expertise is in streaming and sketching algorithms for medical applications. Sketches are structures that allow us to look at big data in a ‘smart’ way. They can significantly compress input while preserving its core properties in order to process big data queries more efficiently. One of our lab’s recent grants has us working on algorithms for wearable devices used to identify classes of arrhythmia in humans.
“Also, we are working on an exciting project funded by DARPA to develop continual Machine Learning (ML) in the setting where the learning agents have small computational power and communication abilities. Together with our collaborators from UT Health, we are applying these methods to improve medical imaging for radiology in the field where standard ML methods will fail. I am excited to see that our algorithms are needed in computational medicine.”
Collaborating with computer scientists in different areas has advantages for Braverman’s students as well. He said his own work grew in depth and breadth when he began interacting on a regular basis with researchers outside his immediate community. Having seen the rich benefits of collaborating with scientists in other areas, he now encourages his own students to explore similar opportunities.
“It may begin with a casual conversation that grows into a joint projector or even a deep friendship. Recently, I’ve been excited by our collaboration with computational biologists, but we can find endless inspiration talking with practitioners in other research areas. It is also humbling to find out how much I don’t know when I meet and begin talking with people in new fields. There is always an opportunity to learn something new when you collaborate beyond your immediate community,” Braverman said.
“In addition to the pure joy of exploring new ideas, it turns out that the other research communities are very welcoming. These are the kinds of connections I want my students to make. Students with expertise across fields become very successful researchers. They develop unique skills in their own field, but they also see how the work applies to other fields.”
Another of his areas of collaboration is the use of algorithms in big data—where the amount of data is massive and computer resources are limited. In one example of such collaboration, algorithms can work in tandem with computer networks to improve outcomes of network monitoring for better security and performance. Additionally, the amount of data used in medical imaging is growing at an explosive pace. With the advent of personalized medicine, more patient data is collected and analyzed in order to customize treatment for a specific person rather than the typical patient with that condition or disease. Braverman’s team hopes to develop algorithms that can help clinicians interpret patient data more effectively as they design personalized treatment plans.
“Algorithmic research overlaps with many different computer science fields,” he said. “When I visited Rice in April, I felt immediate connections with CS faculty members like Professors Lydia Kavraki, Anastasios Kyrillidis, Moshe Vardi, Christopher Jermaine, Todd Treangen, Ang Chen, Eugene Ng, and Anshumali Shrivastava—and those are just a few of the names that come to mind.”
“At least during the twelve years I’ve been here, Rice CS has not had much of a presence in theoretical computer science,” said Chris Jermaine, chair of Rice’s Department of Computer Science. “Vova changes that overnight.”
“Vova does cutting-edge theory in the area of algorithm and data structure design," Jermaine explains. "And he’s passionate about applying his deep understanding of computer algorithms to important problems in machine learning, medicine, and computer systems."
“Machine learning (ML) is one of the areas where we see more and more evidence of algorithms improving efficiency across various areas," said Braverman. "Some ML models have become too large or too expensive to train and this is where new algorithmic thinking can be a tremendous help. That is one of the areas where I hope to work closely with Rice CS faculty colleagues.”
He also hopes to forge a strong relationship with the Texas Medical Center, where he already has ongoing collaborations on medical imaging and radiomics. He said, “My ultimate goal would be to make CS Rice a place where people from different fields and areas come when they need a new algorithmic idea, learn modern algorithmic tools, establish new multidisciplinary algorithmic collaborations or and just have fun discussions at the Duncan Hall or the Coffeehouse.”
“There is fantastic research coming out of Rice. I’d always heard it was an exceptional place, with a beautiful campus, world-class faculty and strong students. After interacting and working with alumni and researchers like Beidi Chen (CS Ph.D. ’20) and my collaborator Stephanie Hicks (STAT Ph.D. ’13), I experienced on a personal level just how special Rice must be.”
Braverman is especially looking forward to doing exciting research with excellent Rice undergrads. His past works resulted in multiple publications on sketching algorithms with undergrad co-authors and he anticipates continuing this tradition at Rice where undergraduate research is particularly strong.
When Braverman began wondering if Rice would be a good fit, he included his family in the decision. He said, “I felt quite at home the moment I stepped onto the Rice campus. After we visited Rice, my family fell in love with Houston and in the end, it was an easy decision to make.”
Prior to joining the Rice University Computer Science Department as a professor, Braverman was an Associate Professor in the Department of Computer Science at Johns Hopkins University with a secondary appointment in the Department of Applied Mathematics and Statistics. Before completing his Ph.D. and M.Sc. at the University of California, Los Angeles, Braverman was a research team lead at HyperRoll, an Israeli startup that has been acquired by Oracle. He also earned M.Sc. and B.Sc. degrees in Computer Science at Ben-Gurion University of the Negev in Israel. In addition to his Rice appointment, Braverman is a Visiting Researcher at Google Research.