The Marconi Society, which promotes awareness of key technology and policy issues in communications and the Internet, honored IEEE Life Fellow Thomas Kailath this month with its Lifetime Achievement Award. It was only the sixth time in the society’s 43-year history that the award has been given. Past recipients include Gordon E. Moore and Claude E. Shannon.
Kailath was recognized for his contributions during the past six decades to areas including information and communication theory, linear systems and control, signal processing, VLSI architectures, smart antenna technology, and semiconductor manufacturing. He also is being honored for his distinctive mentoring of more than 100 doctoral and postdoctoral students, more than half of whom are IEEE Fellows. Among several other major honors, he received the 2007 IEEE Medal of Honor, the organization’s highest award, and the 2012 U.S. National Medal of Science bestowed by former President Barack Obama.
In this interview with The Institute, Kailath, MIT’s first Indian-born student to earn a Ph.D in electrical engineering, discusses his academic contributions, his mentorship of students, and his entrepreneurial ventures.
Congratulations on your award. How do you feel about receiving it?
I am honored and gratified that my work has gained such high recognition from a group of exceptionally distinguished peers. Like all scientists, I don’t do research hoping that there will be an award at the end of it. We do the work for its own reasons and challenges. If it turns out that what we have done is good enough to deserve an award, that is a welcome bonus.
It is important for me to say that after the first decade of largely individual contributions, I realized that a better way to quickly make a major impact on a scientific field—and I entered into a new one roughly every decade—is to work with a group of talented students. Many of my awards recognize us all collectively.
Tell us more about your role as a mentor.
Research is chiefly about asking questions. Some questions are more fruitful than others because they stimulate the right kinds of responses. When working with graduate students as an advisor or mentor, I mostly ask the questions. The students then work hard on them and come back with their thoughts and answers, which I then analyze with them to gain fresh insights and suggest new questions.
With my wider perspective, I try to lead them to the heart of the problem we are attacking. A new result is nice, but one should ask whether simpler and more direct proofs are possible. Are there connections to other results? What is the physical meaning of the mathematical formulas? And so on. In the process, the students begin to ask their own questions. I tell them they’re ready to graduate when they can ask enough of the good questions on their own.
My students say that I’m a tough advisor because I have very high expectations, and they feel considerable pressure to live up to them. But I take comfort from the fact that many of them have successfully moved well beyond their doctoral research areas and are quite distinguished in their fields. More than half are IEEE Fellows, and several have gone on to form companies. It is also gratifying that so many of them are still in close touch with me. Several jointly endowed an annual Kailath Lecture at Stanford, and another former student of mine Guanghan Xu recently endowed a chair at Stanford in my honor.
You’ve also received the National Medal of Science. What was that like?
This medal is, of course, the highest honor any scientist in the United States can receive from the government. And receiving it in the White House from President Barack Obama himself was indeed a memorable experience. The citation itself compactly captured the essence of my efforts: “For transformative contributions to the fields of information and system science, for distinctive and sustained mentoring of young scholars, and for translation of scientific ideas into entrepreneurial ventures that have had a significant impact on industry.”
In the course of an earlier interview with White House staff, I had mentioned that scientists were intrinsically hopeful, and that if we worked hard enough we expected to get good results. In his introduction, Obama brought up this remark, saying that he particularly liked the phrase “intrinsically hopeful,” and he commented that this was true of himself and of the American people.
You were the first Indian-born student to receive a doctorate in electrical engineering from MIT. How did your time at the school influence your career?
Most engineering courses in India in my time depended largely upon memorization and knowing a lot of facts. At MIT, we couldn’t just look up the answers. We had to understand the material and reason our way to the solution.
When I got a scholarship to come to MIT in 1957 to study information theory, I could not have known that I was entering what has been called the Golden Age of information theory. The faculty at the time included Claude Shannon, Robert Fano, Peter Elias, David Huffman and Jack Wozencraft, who were all pioneers in the field. And several of my former classmates, such as Irwin Jacobs, Robert Gallager, Jacob Ziv, Jim Massey, Len Kleinrock, Ivan Sutherland, Elwyn Berlekamp, and David Forney Jr. , are now well-known for their leading contributions to the industry. MIT laid the foundation for my career.
I joined IEEE as a student member. The majority of my research papers, and those of my students, are published in IEEE journals. After graduating in 1961, I spent 15 months in a pioneering digital communications group at the Jet Propulsion Laboratory, in California. In January of 1963 I accepted an offer from Stanford as an associate professor of electrical engineering, and I have been there ever since. My scientific career blossomed at Stanford as the university was rising to its present-day reputation.
You’ve cofounded four companies. Why did you choose to pursue the entrepreneurial path?
I would rather say that entrepreneurship chose me. My work has been largely theoretical, and forming companies was not an ambition I had. But in 1980, my former student Naren Gupta came to me with a proposition for a business idea. We cofounded a company called Integrated Systems, which pioneered software for computer-aided control systems design and later for general embedded systems. The company went public in 1990 and was acquired by Intel.
Not many academics were starting companies in the early 1980s. Things are very different now, and many important companies in Silicon Valley have been founded or cofounded by Stanford and Berkeley faculty. Having tasted success with the first company, I encouraged several of my students to think about how their ideas could lead to new companies. One of the most successful, which went public in 2000, came out of a 1996 Ph.D. thesis on optical lithography, a process used in microfabrication to engrave a pattern on an integrated circuit chip. The more transistors one can put on a chip, the more one can do with the chip. And the number of transistors depends on the width of the lines you can engrave on the chip.
In the mid-1990s it was widely believed that optical technology would run into a fundamental barrier when the lines needed to be thinner than 100 nanometers. I had been encouraged to enter the field of semiconductor manufacturing by Louis Auslander, a program manager in the mathematics department of DARPA [Defense Advanced Research Projects Agency]. However, at the time, he had funding available for research only in manufacturing and not in mathematics. He encouraged us to explore the applicability of tools from control and signal processing to improve any manufacturing problem that we might choose. Our first venture was in the area of rapid thermal processing, and when that effort ended successfully, we took up the challenge faced by optical lithography.
Incorporating ideas from signal processing and communication theory, we showed that the 100-nm barrier could indeed be broken, and decided to form a company, Numerical Technologies, to monetize our methodology. We collaborated with a group at Motorola, which used our software to manufacture a chip with line widths of 90 nm. Once you break a barrier in technology, it’s like Roger Bannister first running a mile in less than four minutes. After that first breakthrough, more and more companies achieved and exceeded that goal, starting with Intel, which first licensed our technology. Now it’s possible to make 7-nm lines using mainly optical lithography. Numerical Technologies was acquired by Synopsis in 2003.
What would you consider your greatest professional achievement?
I am quite proud of my various and varied contributions, including the textbooks and monographs I’ve written and coauthored. As I’ve moved into different fields, I was fortunate that my students and I managed each time to reach the frontiers and make pioneering contributions, often overthrowing the conventional wisdom. Apart from different areas of electrical engineering, I have also done work in probability and statistics, matrix theory, linear algebra, and operator theory. All this went well enough that now I am not only a Fellow of the IEEE but also of the Institute of Mathematical Statistics, SIAM [Society for Industrial and Applied Mathematics], and the American Mathematical Society.
Another source of considerable satisfaction is the wide network of classmates, former students, postdocs, and collaborators that I have been fortunate to build over the past 60-plus years. So, you may appreciate why I cannot single out any particular achievement.