Isaac Chuang, just barely out of Stanford University's graduate school, has become the IBM Almaden Research Center's youngest champion for quantum computing. Quantum effects, such as encoding information on the "spin" direction of a single electron, have the potential to crack tough combinatorial problems that digital computers cannot. But the skills that making computers from quantum effects require call for engineers to cross-fertilize their systems integration skills with principles of physics that are only just being discovered.
Chuang at one time did not believe in quantum computers himself; in 1997 he co-authored a paper in Nature magazine pointing out the unlikelihood of getting a real quantum computer to work in the physical world. Many researchers have published similar papers claiming that no quantum phenomena could be observed in the macroscopic world. The researchers argued that even if quantum computations occurred during picosecond "decoherence" times, the Heisenberg principle spoiled the computation-being able to "observe" a qubit fixed it as a 1 or a 0, rendering it no better than a digital bit.
Fortunately, some physicists and chemists kept insisting that quantum effects would not be so hard to arrange at the atomic level, and, they wondered, what kind of EEs would just give up without trying? That argument got to Chuang, who made it his passion to work with physicists and chemists to debunk the hype surrounding code-cracking quantum computers.
"I surprised myself when we got a quantum computer to work in the real world," Chuang said. "I was collaborating with researchers at the University of California-Berkeley chemistry department, but I expected to fail. However, the answers were there for the looking."
|Isaac Chuang continues to contemplate the state of qubits ever since, to his surprise, he got a quantum computer to work. |
Chuang announced the world's first working quantum computer-a minimal 2-qubit model-in 1998 while still a student at Stanford University, and in 2000 followed with a 5-qubit model forged at his new employer, IBM Almaden, after earning his EE degree from Stanford.
"EEs have something to be very proud of," Chuang said, "but third-millennium engineers will need continuing education in the basic sciences if they are to participate in the invention of next-generation electronics technologies. Despite the power and majesty of VLSI today, when we go to next stage beyond it, we are going to have to go back into the basic science laboratory."
Even though Chuang earned his PhD in Electrical Engineering, he tried to be a generalist in his studies, and in fact took many courses in basic science, physics and chemistry. Had he not, we would all still be sitting on the fence about whether quantum computers were realizable.
"The biggest barrier to inventions today is the broad base and vast amount of knowledge you need in order to invent anything. I can't emphasize enough how important an EE degree is. Electrical engineering teaches you how to think in terms of an overall system and unlike computer science, the EE department tells you how to manipulate that system in the real world. If you want to invent something in the third millennium, an EE degree will be invaluable," said Chuang.
According to Chuang, taking VLSI design to the quantum level will require that systems architectures accommodate the characteristics of the devices being used. In particular, quantum devices are always going to be subject to decoherence-calculations of individual qubits can proceed correctly for short bursts only. To prevent errors, the quantum computer will have to do something like "refresh" qubits while they are in the nebulous 0-and-1 state.
"VLSI today depends on a perfection that is unsurpassed by anything else mankind produces," Chuang said. "We make these very, very large and complex devices with millions of component parts, but so exquisitely perfect that any minute process impurity ruins the entire device. That kind of perfection is unheard of in nature and untenable for the next generation beyond VLSI.
"Nature doesn't do things that way," Chuang continued. "She makes fault-tolerant, robust systems that are capable not only of performing when they are imperfect, but of replicating themselves despite their imperfections. EEs have not been good at that kind of philosophy in engineering, but it's something that must be dealt with before we can be successful with quantum computers. "The paradigm for building quantum computers presupposes that we are using faulty parts, but as we go into a new era of engineering in the third millennium, I think we will finally learn how to build systems from faulty components that nevertheless perform better and are more robust than our ultraperfect VLSI of today," Chuang said.
"If we look around ourselves today, the world is immensely classical, and that is because quantum systems decohere very quickly, subjecting their results to constant errors. The key is to think of decoherence as an error. Then with the proper error-correction techniques built in, we should be able to stabilize these quantum computers and put them to use," said Chuang.
Decoherence is a very strange type of error process that we can overcome only by combining the forces of physics and chemistry with our systemic understanding of computer science and electrical engineering, Chuang noted. Engineers need to take their understanding of how to manage complex electronic systems and apply it to other physical systems, from which quantum computers can be built.