It seems as though quantum computers have gone from the theoretical
pages of some scientist's log book to being practically everywhere in a
very short time. They’re not mainstream yet, or anywhere close, but
increasingly they are finding their way into experimental public-sector
operations.
Cambridge University and Toshiba, for example, are using quantum computing to send completely secure messages around a campus. That works because if anyone tries to snoop the note in transit, it slows it down just enough to destroy the encryption key, leading to nothing but gibberish getting to the other side.
But that's not really quantum computing. That's using the properties of quantum computing to jury-rig an encryption scheme.
Quantum computers on their own are fascinating devices, operating much faster, and somewhat more randomly, than today's supercomputers. The trick in most cases, besides building one, is accurately framing a question, getting the quantum computer to advise on every possible outcome and then picking the ones that get the best results. It's a subtle difference compared with how most traditional computers work. Most computers today are given A and B and told to calculate C. But with a quantum computer, you are more likely to give it C and ask for an efficient way to achieve that with A and B.
The challenge can be in verifying that the A and B given by the quantum computer is the best choice. Often what a quantum computer will do is actually simulated annealing as opposed to hard problem-solving. But again, that depends on the model used, and there is a lot still to learn in this field.
To test quantum computing, NASA and Google are forming the Quantum Artificial Intelligence Lab, which will be housed at NASA’s Ames Research Center. One of the first projects the lab will try to tackle is machine learning, finding out how computers can recognize and learn patterns, and how that might lead to better artificial intelligence.
Hartmut Neven, Google’s director of engineering, wrote in a blog that "Machine learning is highly difficult. It’s what mathematicians call an 'NP-hard' [Non-deterministic Polynomial-time hard] problem. That’s because building a good model is really a creative act."
But machine learning is crucial to better computer models and more accurate predictions, he wrote. “If we want to cure diseases, we need better models of how they develop. If we want to create effective environmental policies, we need better models of what’s happening to our climate.”
The lab’s quantum computer is expected to be operating in the third quarter of this year, and the Universities Space Research Association will invite researchers from around the world to use it. “We actually think quantum machine learning may provide the most creative problem-solving process under the known laws of physics,” Neven wrote.
The computer they will be using for all these creative experiments was built by D-Wave Systems. The New York Times has a good close-up of it, and it looks a bit like something Jules Verne would have put inside one of his ships.
It will be interesting to see if this new quantum computing effort bears any real fruit. Sure, it will act as a baseline and increase our understanding of how the machines work and how to properly program them. But if they will actually deliver any usable hard solutions that could not have been found by a traditional computer remains to be seen.
Cambridge University and Toshiba, for example, are using quantum computing to send completely secure messages around a campus. That works because if anyone tries to snoop the note in transit, it slows it down just enough to destroy the encryption key, leading to nothing but gibberish getting to the other side.
But that's not really quantum computing. That's using the properties of quantum computing to jury-rig an encryption scheme.
Quantum computers on their own are fascinating devices, operating much faster, and somewhat more randomly, than today's supercomputers. The trick in most cases, besides building one, is accurately framing a question, getting the quantum computer to advise on every possible outcome and then picking the ones that get the best results. It's a subtle difference compared with how most traditional computers work. Most computers today are given A and B and told to calculate C. But with a quantum computer, you are more likely to give it C and ask for an efficient way to achieve that with A and B.
The challenge can be in verifying that the A and B given by the quantum computer is the best choice. Often what a quantum computer will do is actually simulated annealing as opposed to hard problem-solving. But again, that depends on the model used, and there is a lot still to learn in this field.
To test quantum computing, NASA and Google are forming the Quantum Artificial Intelligence Lab, which will be housed at NASA’s Ames Research Center. One of the first projects the lab will try to tackle is machine learning, finding out how computers can recognize and learn patterns, and how that might lead to better artificial intelligence.
Hartmut Neven, Google’s director of engineering, wrote in a blog that "Machine learning is highly difficult. It’s what mathematicians call an 'NP-hard' [Non-deterministic Polynomial-time hard] problem. That’s because building a good model is really a creative act."
But machine learning is crucial to better computer models and more accurate predictions, he wrote. “If we want to cure diseases, we need better models of how they develop. If we want to create effective environmental policies, we need better models of what’s happening to our climate.”
The lab’s quantum computer is expected to be operating in the third quarter of this year, and the Universities Space Research Association will invite researchers from around the world to use it. “We actually think quantum machine learning may provide the most creative problem-solving process under the known laws of physics,” Neven wrote.
The computer they will be using for all these creative experiments was built by D-Wave Systems. The New York Times has a good close-up of it, and it looks a bit like something Jules Verne would have put inside one of his ships.
It will be interesting to see if this new quantum computing effort bears any real fruit. Sure, it will act as a baseline and increase our understanding of how the machines work and how to properly program them. But if they will actually deliver any usable hard solutions that could not have been found by a traditional computer remains to be seen.
Posts
