Using entropy quantum computing for solving optimization problems

Question

I would like to follow up on this question (What is Entropy Quantum Computing?).

Company Quantum Computing Inc. announced that they made their quantum computer aimed at solving binary optimization problems available to public. They claimed that the computer uses what they call entropy quantum computing:

Natural quantum states interact freely, influencing and impacting each other as they evolve and change. This natural interaction significantly impacts the accuracy and scale of first generation NISQ computers. The expense of hyper-cooled, vacuum environments doesn't eliminate loss of information, significant errors and limited scaleQCI's Entropy Quantum Computer, Dirac, harnesses the true fundamentals of quantum physics to overcome these limitations. It operates on open quantum systems, carefully coupling a quantum system to an engineered environment, so that its quantum state is collapsed to represent a problem’s desirable solution.

I understand that they somehow program an engineered environment to represent an optimization task and then evolve a quantum state to the ground state of the environment. This seems similar to quantum annealers principle. However, how the noise is reduced with help of the engineered environment is a little bit opaque.

I also found this description of the computer principle of function:

The company indicates the devices uses backaction with the environment to evolve the quantum system into a coherence-free sub-space.

It sounds to me like Maxwell's demon or Perpetuum mobile - they use noise of the surrouding environment to reduce noise in the quantum computer.

Moreover, I find out that the computer somehow exploits photonic technology. So, it seems to me that the entropy quantum computing is in fact photonic-based quantum annealing. Does it make sense?

I could not find any paper on the entropy quantum computing, the only relevant literature is question on this site I linked above. Any link to paper(s) would be highly appreciated.

Answer 1

Based on advice by gIS I looked for quantum reservoir engineering. In short, it is a technique allowing to stabilize certain quantum state $\mathcal{S}$ with engineered coupling to an environment $\mathcal{E}$ (sometimes called reservoir). A ground state of multi-system $\mathcal{S-E}$ is such that allows quantum state $\mathcal{S}$ to be stable. Or in other words, to preserve coherence of superposition for "long(er) time". Very theoretical background is provided in "eponymous" article Quantum Reservoir Engineering.

It seems to me that such configuration can be likened to a fridge. With help of compressor (the coupling) a heat from fridge (the noise influencing state $\mathcal{S}$) is pumped out to surrounding environment $\mathcal{E}$. This process is not in contradiction with $2^\text{nd}$ law of thermodynamics as work is carried out to cool the inside of the fridge. Similarly, the above described process of stabilizing quantum state requires external work to establish the coupling and transfer noise to the environment.

I found two practical application of the quantum reservoir engineering. First one used for resetting qubit to state $|0\rangle$ (Quantum reservoir engineering and single qubit cooling) and second one for stabilizing Bell state (Stabilizing a Bell state of two superconducting qubits by dissipation engineering). Similarly, it should be possible to stabilize any quantum state with proper coupling to proper environment (the environment can also be "engineered").

Quantum Computing Inc. claims that its quantum processor can be used for solving optimization task. This would mean that the processor is similar to quantum annealers. Based on the information above, I would infer the function of the processor as follows. The processor is set to initial state (probably uniformly distributed superposition as in case of annealers or VQE/QAOA algs.) and coupled to the environment. After that the coupling is changed so that it reflects a solved task. In the end, the processor contains stabilized ground state of coupling Hamiltonian describing the solved task.

I am apologize if this is crude description, however, there is not so much information available. If anybody could add anything or make my answer clearer or more accurate, please do so.