Yesterday was September 1, 2020 and we zoom-recorded the first session of 47-779, Quantum Integer Programming.
This course is also being taught at IIT-Madras, with one week offset, using these zoom-recordings: ID 5840, Quantum Integer Programming.
As we finished the class, my mind went to Come September, and to one of its memorable lines, by Gina Lollobrigida (playing Lisa Fellini):
I don’t have to make sense. I am Italian!
What is the image that your mind conjures up when someone says quantum computer?
Something large, perhaps the size of a room, with its qubits being supercooled and using up a lot of energy, sitting somewhere remote, too expensive to buy one for personal use, and accessible as a cloud service.
Indeed, Amazon Braket is one such cloud accessible platform from which we can access D-Wave, ionQ and Rigetti machines.
Thanks to a NSF grant, and collaboration from Amazon, the Quantum Computing Group at Tepper now has access to Braket with funds to access these quantum computers to do research on Quantum Integer Programming.
It is a wonderful platform, and we are enjoying exploring it.
I wondered, using my maximally inverse thinking:
Is it possible to have a quantum computer in our pocket, in room temperature, using low amount of energy, that is affordable at a retail level, constructed from off-the-shelf parts and mature technologies?
Crazy? Not quite!
Fusing George Bernard Shaw with Oscar Wilde to obtain The Importance of Being Unreasonable 😏:
The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man.
Let me introduce you to the follow up of Book One:
Neo-Quantum Organon, Book Two: Chip-based Ising Computing Machine (CICM).
An Ising Computing Machine (ICM) obtains the desired solution to a combinatorial optimization problem through collective state computing rather than sequential state computing.
ICM relies on the Ising model that is known to be Turing-complete, meaning that it can solve any classical computing problem.
ICM is analog in nature, intrinsically offering an entirely different type of parallel computing than digital computing devices.
ICM can be made using superconducting Josephson junctions, such as the D-Wave machine, but this requires liquid helium cooling (temperature less than 15 milliKelvins), is the size of a room, has 2048 qubits, limited connectivity (and so is not a Universal Ising solver, 5600 couplings, which have limited fidelity), requires 25kW of power to operate and retails for over USD $15 million.
Chip-based ICM (CICM) that we are envisioning operates at room temperature, is of size 10 square centimeters (or less), has power consumption of less than 10 w, with 10000+ qubits.
Two possibilities (there are actually many more) for CICM are:
CICM-OPO (Optical Parametric Oscillator)
CICM-MTJ (Magnetic Tunneling Junction)
The key aspects of CICM are: (a) how many qubits (b) how are they connected and (c) what is their speed of evolving?
Different CICMs have different advantages and disadvantages with respect to the above three aspects, and so are likely best suited for different applications. For example, with CICM-OPO, the information is encoded in the phase of the photon, and so its speed is very high but how many different qubits can be physically placed on a small area of the chip is limited.
Thus, it is worthwhile to fabricate multiple CICM-types rather than focus on just one.
The good news is that the design and fabrication of these CICMs share many common activities and technologies, and furthermore, progress or hindrances in one type can inform another. For this reason, our group of international collaborators (US, India, Singapore, Japan) is pursuing the design, fabrication and testing of three CICMs.
Where is the clean room? Where is the fab? We have support from and are collaborating with:
Global Foundries and Broadcom.
One can frame this unreasonableness of ours, a merry band of quantum researchers, as a modern-day underdog situation:
CICM v Superconducting Qubits
CMU, IITM, NUS v IBM, Google, Amazon
David v Goliath
Malcolm Gladwell, however, has a different interpretation of David v Goliath, as you may heard in his TED Talk (talking about the sling):
It’s not a child’s toy. It’s in fact an incredibly devastating weapon. … If you do the calculations on the ballistics, on the stopping power of the rock fired from David’s sling, it’s roughly equal to the stopping power of a [.45 caliber] handgun. This is an incredibly devastating weapon. … When David lines up … he has every intention and every expectation of being able to hit Goliath at his most vulnerable spot between his eyes.
As a cinephile, this reminds me of Harrison Ford, and I would say:
It’s like Indiana Jones shooting the swordsman in Raiders of the Lost Ark.
Putting it in the form of a well-known maxim:
Don’t bring a knife to a gun fight.
Of course, one has to incorporate game theory appropriately, think two steps ahead, something that Sean Connery did not do in The Untouchables, in one of the scenes that hurt so much: