# Use of physical environment to obtain useful results

I dare to ask a broad layman question expecting a layman simple answer. I'm no scientist in any field.

Question: Does the following concept make sense or does it exist and actually being used? > Given some initial (digital) data, it would in some cases be possible AND more efficient to cast these "outside" of the digital computer (most probably in a form of some wave), letting it be altered through specifically fitted physical environment and receiving them back as digital data in order to solve some problem/obtain useful result rather than processing those initial data digitally by the software function. Time spent on creation of this "physical filter" (most likely not solid, e.g. magnetic field) should most probably be added to the overall processing time consumption for it to make sense. Also: it would be expected not to use this method to obtain very exact answers to exact problems, but rather to draw some useful data from some complex initial data. The filter would probably not be static, but could change according to required setup.

An illustrative example bellow (not actually feasible)

problem: Is value of x higher than any of the values in given very large set?

software solution: function iterating through each value in set and comparing value of x to it. When none is larger or equal when the cycle ends, x is higher than all, otherwise x is not larger than all and cycle can break early.

solution using physical filter: filter has the form of opaque columns of height that represent values in given set. Value of x is outputted in a form of a light ray at corresponding height "through" the filter. When no column is higher or equal, light is received by the sensor on other side.

When not considering the time required to construct such filter, the solution to the problem using the filter should be much faster than software solution, no?

Hope the question is not too stupid or too irrelevant to this stack.

• This is not a stupid question at all! In fact, this is likely how the first practical quantum computers are going to be implemented: a classical computer will "control" (i.e. give input to and take output from) the physical process that makes up the quantum computation. – Mikhail Rudoy Dec 16 '17 at 6:35
• Gaudi, famous architect, known especially for his La Sagrada Familia cathedral, used very similar method, but in different physical settings. When he have to compute optimal roof shape he used a system of ropes and weights, to simulate real building static forces. And then just by measuring rope shapes and forces, he create real building. It was at times when no computer was available for him, and nobody has stupid idea the universe is digital simulation... – kakaz Dec 17 '17 at 9:58
• @MikhailRudoy thank you, yes, that really does seem to fit in my description. On the other hand, to my understanding, it is something that is still very experimental and difficult to set up. I was thinking maybe the concept could be used in a simpler more managable scenerio as well. – Demo Dec 17 '17 at 15:59
• There is the theoretical idea of "Boson sampling" to use a special quantum mechanic effect to solve an ${\#P}$-complete problem (I'm no expert on the field). – Sudix Aug 31 at 5:28
• An important mathematical operation that is performed in many programs is the Fourier transformation. Typically this is done by using the Fast Fourier Transform algorithm whose runtime scales with $\mathcal{O}(n \cdot \text{log}(n))$. There are at least ideas to construct a hardware coprocessor that performs such calculations by making use of optics. With such an approach the scaling can be reduced to $\mathcal{O}(n)$. See e.g. nature.com/articles/s41598-017-13733-1 – Gregor Michalicek Aug 31 at 15:58