Welcome to Machine Elegant. On this blog, I plan to discuss topics of a technical nature, including news articles, projects and big ideas. And now for some news...
Engineering professor Nader Engheta and his team have demonstrated a metamaterial device that can function as an analog computer, validating an earlier theory about 'photonic calculus.'
We are living in exciting times for anyone who tracks technological progress. We are making rapid strides on all fronts and the pace of progress is being accelerated by recent progress. An obvious example is silicon-based computer technology which has created a "virtuous circle" whereby advances in computers are being used to design and build the next generation of computers. But this same phenomenon is occurring in many domains and it is beginning to "cross-breed", that is, advances in traditionally unrelated domains are beginning to produce speed-ups in one another.
Quantum computing (QC) is getting a lot of attention right now as the next up-and-coming technology. But without some kind of major breakthrough in materials science, we may be calling the game too early on QC. The problem is, quite simply, noise. From the perspective of a very cold object -- say, a chunk of frozen nitrogen -- we live in a very hot universe. And physical heat causes noise. QC requires each qubit to be kept in near-perfect isolation from its physical environment because quantum states are maximally sensitive to noise. Interference from the external environment causes the quantum system to become entangled with the external environment, which corrupts the state of the ongoing quantum computation.
Metamaterials are remarkable in that they allow us to take advantage of the physics of microscopic systems (such as photons), but at macroscopic scales. This up-scaling property allows us to build metamaterial systems that are more or less impervious to the effects of noise at room temperature. To be clear, I am not suggesting that a metamaterial photonic computer can perform general-purpose quantum computation. But we might just be able to utilize photonic computing as a bridge technology between the state-of-the-art silicon computers, and quantum computers.
How?
Quantum computers, when we build them, will not replace traditional, digital computers because (a) quantum computers are good for fundamentally different kinds of computations than digital computers and (b) digital computers are basically impervious to noise and quantum computers will never be. The kinds of computations that quantum computers excel at (by comparison to digital computers) are exactly the same kinds of equations which matamaterial photonic computers can also perform efficiently. Take quantum machine learning (ML), for example. We know that a QC would massively speed up ML tasks vis-a-vis digital computers. But, for a given ML task, a metamaterial photonic computer could provide exactly the same speed-up, allowing us to solve massive problems that are untouchable by digital computers in the blink of an eye. The disadvantage, of course, is that the metamaterial photonic computer is not (yet) configurable. So, it's a bit like burning a DVD or fabricating a silicon circuit. Write-once, use-many.
The potential of QC is so immense that it is worthy of large investment. But, as we all know, large investment doesn't guarantee success. I keep an eye out for "alternative computing" technologies like quantum dot cellular automata and metamaterial photonic computing because I believe we need to hedge our bets on QC. Even though these computing paradigms are not as powerful as QC, they are orders of magnitude more powerful than digital computers on very important problem domains (such as ML) and may help us solve the herculean task of building a working, at-scale quantum computer.
Engineering professor Nader Engheta and his team have demonstrated a metamaterial device that can function as an analog computer, validating an earlier theory about 'photonic calculus.'
We are living in exciting times for anyone who tracks technological progress. We are making rapid strides on all fronts and the pace of progress is being accelerated by recent progress. An obvious example is silicon-based computer technology which has created a "virtuous circle" whereby advances in computers are being used to design and build the next generation of computers. But this same phenomenon is occurring in many domains and it is beginning to "cross-breed", that is, advances in traditionally unrelated domains are beginning to produce speed-ups in one another.
Quantum computing (QC) is getting a lot of attention right now as the next up-and-coming technology. But without some kind of major breakthrough in materials science, we may be calling the game too early on QC. The problem is, quite simply, noise. From the perspective of a very cold object -- say, a chunk of frozen nitrogen -- we live in a very hot universe. And physical heat causes noise. QC requires each qubit to be kept in near-perfect isolation from its physical environment because quantum states are maximally sensitive to noise. Interference from the external environment causes the quantum system to become entangled with the external environment, which corrupts the state of the ongoing quantum computation.
Metamaterials are remarkable in that they allow us to take advantage of the physics of microscopic systems (such as photons), but at macroscopic scales. This up-scaling property allows us to build metamaterial systems that are more or less impervious to the effects of noise at room temperature. To be clear, I am not suggesting that a metamaterial photonic computer can perform general-purpose quantum computation. But we might just be able to utilize photonic computing as a bridge technology between the state-of-the-art silicon computers, and quantum computers.
How?
Quantum computers, when we build them, will not replace traditional, digital computers because (a) quantum computers are good for fundamentally different kinds of computations than digital computers and (b) digital computers are basically impervious to noise and quantum computers will never be. The kinds of computations that quantum computers excel at (by comparison to digital computers) are exactly the same kinds of equations which matamaterial photonic computers can also perform efficiently. Take quantum machine learning (ML), for example. We know that a QC would massively speed up ML tasks vis-a-vis digital computers. But, for a given ML task, a metamaterial photonic computer could provide exactly the same speed-up, allowing us to solve massive problems that are untouchable by digital computers in the blink of an eye. The disadvantage, of course, is that the metamaterial photonic computer is not (yet) configurable. So, it's a bit like burning a DVD or fabricating a silicon circuit. Write-once, use-many.
The potential of QC is so immense that it is worthy of large investment. But, as we all know, large investment doesn't guarantee success. I keep an eye out for "alternative computing" technologies like quantum dot cellular automata and metamaterial photonic computing because I believe we need to hedge our bets on QC. Even though these computing paradigms are not as powerful as QC, they are orders of magnitude more powerful than digital computers on very important problem domains (such as ML) and may help us solve the herculean task of building a working, at-scale quantum computer.
Comments
Post a Comment