Imagine we have the following equation: y=xz. We have y but not other ones. Note that y is like a matrix and we could as many sample we want. It is the values obtained from sensors. This means it will be a m*n dataset where m represents number of samples and n represents the number of sensors. I am wondering whether we can use machine learning techniques (any method including GAN) to solve the equation and get at least one of the variables of x or z. Thanks.

  • 2
    $\begingroup$ It's ridiculously easy to find a solution: pick the identity matrix for x and y for z. $\endgroup$ – Erwan Feb 24 at 19:39
  • $\begingroup$ Could you please give a toy example. We just have and don't have x and z. $\endgroup$ – Arkan Feb 24 at 21:13

Machine Learning is useful for problems which can't be solved (or not efficiently) with a deterministic method. The problem you propose is extremely simple since it doesn't even require any calculation, so there is no point using ML for it.

Could you please give a toy example. We just have y and don't have x and z


$y=\begin{bmatrix}1 & 2\\ 3 & 4\\ 5 & 6\\ 7 & 8\\ \end{bmatrix}$

An easy solution is to define $x$ and $z$ as follows:

$x=y=\begin{bmatrix}1 & 2\\ 3 & 4\\ 5 & 6\\ 7 & 8\\ \end{bmatrix}$ and $z=\begin{bmatrix}1 & 0\\ 0 & 1\\ \end{bmatrix}$

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.