Last week, Google’s AI team announced that its tests at NASA’s Ames Research Center on the D-Wave 2X quantum computer yielded exceedingly pleasing results: it was 100.000.000 times faster at solving the problems set to it than a conventional computer using quantum annealing as opposed to the simulated annealing conventionally used for such problems.
Yes, you read that right. Google partnered with NASA, a consortium of universities (the Universities Space Research Association – USRA) and with D-Wave Systems Inc. in 2013 so that the first three organizations could test and refine the latter company’s quantum computers to further the development of AI and space travel.
Since its Orion prototype quantum computer, presented around 2007 D-Wave has gradually increased the technology’s capabilities. If the Orion prototype worked on a 16-qubit processor, the current D-Wave 2X works on one with more than 1000 qubits.
The way quantum computers are supposed to work is that, because of the strange phenomena that take place at a quantum level, qubits can have both a 0 and 1 value as well as an additional 0 and 1 at the same time value, due to superposition.
Since conventional processors operate with just 2 alternative values (1 or 0), quantum computers should theoretically be way faster.
And judging by Google’s announcement last week, the theory proved out as their AI team demonstrated these incredible speeds using the D-Wave 2X quantum computer.
Or did they? Because here’s the catch.
Ever since its first claims at having built a functional quantum computer, D-Wave has been consistently surrounded both by awe, excitement and praise, as well as by criticism, accusations and dismissal by respected scientists.
The first camp is pretty clearly defined by the Google, NASA and USRA scientists working with the computers and other non-affiliated ones.
But the second one shouldn’t be neglected either, boasting prominent names such as Umesh Vazirani (one of the founders of the theory of quantum complexity), Wim van Dam of UC Santa Barbara and Scott Aaronson of MIT, the latter proclaiming himself the “Chief D-Wave Skeptick” (with a short hiatus) for quite some time.
Making matters more confusing is that an article published in Nature in 2011, does support D-Wave’s technological claims, stating that the company’s chips do show some of the quantum mechanical properties required for quantum computing.
But does that mean that D-Wave computers actually operate on a quantum level? Wim van Dam and others believe there is no way to tell for sure at the moment. And the more hardcore critics like Varizani or Aaronson say that D-Wave misunderstood some quantum principles, that its claimed speedup flat out isn’t true and that they inflate their results.
Still, Google claims that the results are real and really quantum. And hints at future applications like fast and optimized space travel plotting, drug testing, airport coordination, encryption and many others.
Time will tell. What do you think?
Image source: 1.
