As an undergraduate at Harvard, Daniel Gottesman (pictured left) was interested in physics, but he never liked doing experiments. "What I really didn't like was the data?it's never as clean as you would want. You have to do analysis, and that stuff I don't like," he says.
Gottesman's distaste for data analysis led him to a career in theoretical physics, a field in which some 15 years later he is widely regarded as a rising star. According to John Preskill, MacArthur professor of theoretical physics at Caltech and Gottesman's Ph.D. thesis advisor, Gottesman knows how to pick projects. "What makes him special is his ability to judge which problems are the best to work on. Perhaps the most important skill for a researcher is being able to identify which problems you have a good chance of solving and which solutions would be most important," Preskill says.
After joining Preskill's group in 1992, Gottesman did a paper on black hole evaporation, but "I wasn't getting very far," he recalls. Then Peter Shor, who was at AT&T at the time (he's now at MIT), published an algorithm for factoring large numbers on a quantum computer. "It was extremely significant," recalls Gottesman. "If you can factor large numbers you can break the RSA [Rivest, Shamir, and Adleman] public cryptography system, which is used everywhere, including secure Web applications." More significantly, from his point of view, the algorithm demonstrated that a quantum computer really can be faster than a classical computer.
Existing in a range of different states
A quantum computer would operate by taking advantage of the fact that a particle can exist simultaneously in a range of different states. Instead of being restricted to a series of 1s and 0s, as in a classical computer bit, a quantum computer's qubit is both a 1 and a 0, and many states in between. In effect, the quantum computer can process all of these states simultaneously, giving it staggering computational power ? at least in theory. The catch is that as soon as you observe the quantum states, they collapse (or decohere) into a random state.
Because of that limitation, it wasn't clear that a quantum computer could ever out-perform a classical computer. Shor's algorithm got around the problem by using different branches of observation and allowing them to interact with each other to produce a superimposed interference pattern, which could be observed without disrupting the quantum states. "It allows you to extract some information about the global state without measuring any one of them," says Gottesman.
Still, quantum computers were vulnerable to random collapse even in the absence of external observation. A simple interaction with its environment could cause a qubit to decohere, causing an error that could propagate and drag the quantum computer to a standstill. That's where Gottesman came in. Preskill suggested he look into the problem, and he responded by developing a mathematical framework to describe quantum error correction codes. The framework "makes it easier to think up new codes and understand how they behave," says Gottesman.
Although it was done under his supervision, Preskill insists that the work is primarily Gottesman's. "When he first started to explain it to me, it was hard for me to understand because it was so innovative."
The work made a name for Gottesman and continues to have a big impact on the field. "[He] has helped to lay the mathematical foundation for a class of quantum error correction codes that will likely form the basis of future quantum information technology," says Jeremy Levy, associate professor of physics and astronomy at the University of Pittsburgh, and director of the Center for Oxide-Semiconductor Materials for Quantum Computation.
After receiving his Ph.D. in 1997, Gottesman went to Los Alamos for a postdoc because it was a "powerhouse" of quantum computation, he says. After a year, he began to search for tenure-track academic positions, but he could find none for quantum computation in North America, possibly because quantum computation does not fit into the mandate of any particular department, he says. Everyone working on it at the time had switched from another field.
With no academic positions available, he cast about for another postdoc, and was favorably impressed by the camaraderie and intellectual rigor of Microsoft's Theory group, which consisted of theoretical computer scientists and mathematicians working on research with potential application to future Microsoft products.
Gottesman continued to look for faculty positions, but there were still no openings. Then, in 2000, a couple of companies interested in starting a quantum computing research program offered him positions. At the same time, the Clay Math Institute (CMI) offered him its prestigious prize. It was a difficult decision. He could take the CMI prize anywhere, but it lasted only two years and would force him back into a job search. On the other hand, the tech bubble was in the process of bursting. "Companies may like to support research when there is a favorable business climate, but if things are less rosy, they may cut back or try to shift to a more short-term focus? he danger of that was obvious," he says.
He also felt it was time to go back to an academic environment. "At both Los Alamos and Microsoft, I had felt that basic research was really just a minor sideline to the institutions' main mission--nuclear weapons in the case of Los Alamos, software in the case of Microsoft--and I decided I wanted to be someplace where the institution's long-term goals were aligned with mine. Also, the symmetry appealed to me of doing postdocs in a government research lab, an industrial research lab, and a university, in groups with a mix of physics, math, and computer science emphasis." So he took the CMI fellowship and went to Berkeley for another postdoc, in the computer science department.
Quantum states to prevent eavesdropping Throughout the moves, Gottesman continued to work on quantum error correction, but he also moved into the field of quantum cryptography, which uses quantum states to prevent eavesdropping. If someone listens in on a message encrypted with a quantum key, it causes a collapse of the quantum state and leads to an error that can be detected on the receiving end. Such a coding system is theoretically unbreakable.
In fact, the two fields are related. "In order for a quantum error-correcting code to work, it must prevent the environment from learning about the encoded quantum state?because measuring a quantum state will necessarily disturb it. This built-in secrecy makes quantum error-correcting codes ideal for quantum cryptography," says Gottesman.
Near the end of his time in Berkeley, academic jobs finally began to open up. As he was considering a couple of offers, several colleagues forwarded him an ad for a position at Waterloo, Ontario-based Perimeter Institute, a private non-profit institute dedicated to theoretical physics. "I was attracted by the set-up of an independent research institute dedicated to top-quality research and by the commitment to build a strong quantum computation group."
He seems poised to continue his success. Hoi-Kwong Lo, an associate professor in the department of physics and electrical and computing engineering at the University of Toronto, also studied under Preskill at Caltech, leaving a couple of years ahead of Gottesman. The two began collaborating while Gottesman was still in graduate school. "I have always been extremely impressed by the clarity of his thoughts. In my opinion, this is peerless."