Re: Inversion of an algorithm



Daniel Pitts wrote:
Jean Guillaume Pyraksos wrote:
Hi !
Say I have a one-one function f from a set A to a set B. I have
an algorithm to compute f.
Has anyone worked on getting *automatically* an algorithm for the
inverse function from B to A ? Or is there a deep result about that ?
Has this problem something to do with automatic differentiation ?

f is *not* specifically a math function, but a recursive one on trees.

Thanks for any pointer,

JG
Not all functions, let alone algorithms, have an inverse. Therefor it is impossible to find it automatically.
....

Remember the OP restricted the discussion to one-one functions, and
subsequent discussions effectively limited it to bijections - the
inverse is required only where it exists.

Patricia
.


Quantcast