On Feb. 1, you’ll find Distinguished Speaker Jeffrey D. Ullman in Donald Bren Hall 6011 at 11 a.m. giving his talk, “Data Science: Is It Real?” The S.W. Ascherman Professor of Engineering (Emeritus) from Stanford University and CEO of Gradiance Corp. will discuss how different communities approach data science.
“He’s a giant figure, and students should appreciate this opportunity to see and hear him,” says Computer Science Professor Dan Hirschberg. “They will be motivated and inspired.” Hirschberg speaks from experience, having been advised by Ullman in the 1970s while earning his Ph.D. at Princeton University. Ullman also earned his Ph.D. from Princeton and joined the faculty in 1969.
In 1979, Ullman moved to Stanford University, which is where Computer Science Professor Chen Li had him as his Ph.D. adviser. “It’s a very rare opportunity to see this legendary researcher, who has had a big influence on the CS community,” asserts Li.
Hirschberg and Li both agree that Ullman is friendly and approachable, yet they warn students to expect honest and direct responses to any questions asked. “He can be tough and will express his opinions in public very frankly,” says Li. “People should be prepared for that kind of honesty in his talk.”
“In a university environment, one should have frankness and openness,” adds Hirschberg — especially coming from a researcher of Ullman’s caliber. “He has excelled in so many different areas of computer science. His books are classics in all these different directions.”
Ullman has written about everything from database systems and data mining to compilers, automata theory and algorithms. Furthermore, he has received the SIGMOD Contributions Award (1996), the ACM Karl V. Karlstrom Outstanding Educator Award (1998), the Knuth Prize (2000), the SIGMOD E.F. Codd Innovations award (2006), the IEEE John von Neumann Medal (2010) and the NEC C&C Foundation Prize (2017).
He is well-respected in the field and has advised countless students, including UCI Computer Science Professor Emeritus George Lueker, who earned his Ph.D. from Princeton in 1975. “Ullman was an excellent adviser,” says Lueker. “He was intellectually generous and gave his students space to work independently.”
Li agrees, saying Ullman “was fun to work with, and his style was more hands-free,” so students had a lot of leeway in conducting their research. “Because he gave students freedom, he had quite a few legendary students, such as Google founder Sergey Brin and even before that, the founders of Junglee.”
Hirschberg, Li and Lueker all three recall their adviser’s wry humor. Hirschberg’s copy of “A Professorial Chrestomathy” includes some of Ullman’s quotes, such as “Aha! I’ve again covered up the fact that I don’t know what I’m talking about with a laugh.”
On a more personal note, Li explains that Ullman helped ease his transition from China to the U.S. “It took quite a bit of effort to get used to the language, the culture and the graduate program.” Ullman helped him better grasp the language through detailed edits of writing assignments and suggested stylistic choices that enhanced his arguments. For example, Ullman once told Li not to overexplain things. “I spent too much time justifying why I made an assumption, and then he told me, ‘strategically, you don’t need to do it.’ He said ‘it makes you look suspicious.’” Li says this strategy “extends well beyond the scope of technical writing.”
On the research side, Li says Ullman also taught him the importance of taking risks: “It’s okay if you fail. Just try some things and see which one works out.” This philosophy encouraged him to branch out in his research.
Ullman also “emphasized the importance of solving real problems,” says Li, “not just artificial problems.” Hirschberg similarly notes that “the flavor of the things I’ve done has been theoretically oriented but with more than a nod toward usefulness.” For example, Hirschberg’s thesis covered the longest common subsequence problem, which subsequently was used to compare DNA sequences.
When you look at the almost five decades of Ullman’s research, his impact is clear. “He started with automata, compilers, data structures, algorithms and VLSI, and then in the late 1980s or early 1990s, he switched to the area of databases,” says Li. “He is a big believer of the importance of the data management side…. I believe what he is going to advocate for [in his talk] is the importance of the more traditional data side.”
Specifically, Ullman will discuss “the various ways in which data science is approached by different communities [and] some consequences of these approaches.” In his example of why data science is not machine learning, he will “sketch two important algorithms not in this class: locality-sensitive hashing and approximate counting.”
Ullman’s talk is part of the Computer Science Department’s Distinguished Lecture Series, which is free and open to the public.
Lueker, who retired from UCI in 2009, plans to be back on campus with Hirschberg and Li to hear the talk. “He’s a very entertaining speaker,” says Lueker. “I’m definitely looking forward to his talk.”
— Shani Murray