Hardware Initiative at Quantum Artificial Intelligence Lab

Posted by Hartmut Neven, Director of Engineering

The Quantum Artificial Intelligence team at Google is launching a hardware initiative to design and build new quantum information processors based on superconducting electronics. We are pleased to announce that John Martinis and his team at UC Santa Barbara will join Google in this initiative. John and his group have made great strides in building superconducting quantum electronic components of very high fidelity. He recently was awarded the London Prize recognizing him for his pioneering advances in quantum control and quantum information processing. With an integrated hardware group the Quantum AI team will now be able to implement and test new designs for quantum optimization and inference processors based on recent theoretical insights as well as our learnings from the D-Wave quantum annealing architecture. We will continue to collaborate with D-Wave scientists and to experiment with the “Vesuvius” machine at NASA Ames which will be upgraded to a 1000 qubit “Washington” processor.
Nuno Edgar Fernandes's profile photoJUNIOR FERREIRA's profile photoBrendan Sterley's profile photoKirigaya Suguya's profile photo
Exciting. Sounds like the research is expanding in scope, too.
Does this make Vesuvius the fastest, man-made, processing unit? Or just one that can perform an incredible amount of parallel calculations at once?
Wow, it's wonderful to hear, keep the good news coming (:
Quantum AI is gonna create epic strides in humanities progression. However... Not if humanity does not progress themselves consciously.

Oh man, QAI video gaming IS GONNA BE THE SHIT
Where does someone get the booklet with enough information to learn and know about this???? I checked in my happy meal and in my cherioos and couldnt find them u u 
Too cool! Huge strides already, it's going to be an Intresting couple years. Ugh.. I'm going to have to learn all new languages... 
when did a response become agitated by its reply
The term 'quantum AI' is as significantly meaningful as Heisenberg's principle of certainty.
How does a Research Scientist with QC Physics background apply for a job in the Google Quantum A.I. Lab Group?
Scott Aaronsons blog calls this announcement 'B.S.'
Will this new chip architecture aim to implement the Quantum Gate model or stick to Quantum annealing?
Boris Epshtein, how do you define 'working'?  The D-Wave machine is designed to do quantum annealing, so far it looks to do exactly that. But I  hear form people who code for the machine that they would like to have hardware based error correction. 

As the number of qubits increases this will become more pressing.  From the press releases it's not clear to me if this effort is aimed at getting a better annealing chip with error correction, or if this goes straight for the moon shot i.e. gate based QC?
Hurry up and take over the world already, Google(all heil my new overlord).
 I for one  welcome our new AI overlords  and their infinite knowledge  all hail Google AI
For those looking for more information on the field in general, "Mike & Ike" is considered one of the key references - it was the text I used it when I took Quantum Computation in my MSCS program.  Michael Nielsen, Isaac Chuang.  Quantum Computation and Quantum Information, Cambridge University Press (10th Ed. pub. 2011).  
This is annealing with surface code error correction. The qubits are in a 2D grid where nearest neighbor entanglement is used for error correction. The substrate is AlSa but otherwise its DWave architecture.
+Henning Dekant They are just looking to increase coherance time to the point the spectral gap is closed enough to see effective annealing
in other words, Google bought a DWave, played with it, studied it, learned from it, and now will build their own new and improved HW. makes sense
So what DWave built was the equivalent of the ENIAC, actually earlier as we could call DWave an analog computer in essence. The quest for the quantum gate producing a true 'digital' quantum computer is the goal and seemingly still a long way off. 
Hm, I think your observation may be of the square peg/round hole variety.  :-)  Exploiting quantum superposition for computation is analogous to the challenge of interfacing a digital computer to the analog nature of many problems.  In order to make use of the quantum computation's results, we need to translate it into something we already know how to deal with, whether that is a digital or an analog result.  But I'd say that in effect, we have three approaches: analog, digital and quantum.  

BTW, if you're implying a chronological sequence, analog --> digital, I'd suggest you take a look at recent work done in AI employing analog computing elements.  Analog computing never went away.  :-)  
+Ian King The problem is, with blurry vision squares look round. Allow me to build you bifocals: Analog and digital computing refer to both a process and an architecture. Quantum refers only to an architecture and not a 'process'. There is no quantum 'way' to  compute. Annealing is a process of finding local minima by maintaining coherence through the spectral gap, terminating the transverse field and allowing the adiabatic process to equalize. There is nothing inherently 'quantum' about it, in fact the problem with DWave was that it could be simulated on a classical computer and that simulation was faster. The goal is a true digital quantum gate, not a 'quantum' way of computing, whatever that is. Its true that the qubits (J-junction SQUIDS) are coupled to traditional A/D converters for final processing, as we do taking analog reading to digital, but annealing is an analog process.
I'd love to look at your analog AI, but you are a little thin on references or links.
+J. Neville Groff D-Wave is not allowing universal adiabatic computing by restricting the Hamiltonian to the Ising model but the former has been shown to be computationally equivalent to the gate model:  http://arxiv.org/abs/quant-ph/0405098

The gate model is not the only one in town although there certainly is quite a bit of myopia when it comes to alternatives (another one would be quantum cellular automatons for instance).
+Henning Dekant Amazing. I'm not in the field and I learned the math indirectly through studying category theory but that document looks foundational judging by date. It looks to be the basis of DWave and through to the current 'surface code' of the new machine? It looks like they describe the 2D qubit grid proposed in the new Martinis machine in theorem 1.3. From the description of the gate equivalent operations in that paper, specifically the 3 qubit reversible gate that seems to imply a universal Toffoli gate, it looks like Martinis machine is not just a path to longer coherence and better annealing but to actual gate-equivalent computation. Well, I guess when you combine superposition, reversible universal gates, entanglement and other operations that would render the operation of a quantum computer as a whole unsimulatable on a classical system then you really do have a 'quantum way' of computing.
Hdhdu b cdidbd7ducb ichd dd I I hcicudnd jc cidjd7d
Dkd c icjjdjdj

Add a comment...