Two dimensional PN functions
- Already Purchased? View Article
- Subscription Options Learn More
There is a differential operator ∂ mapping 1D functions φ:G→C to 2D functions ∂φ: G×G→C which are coboundaries, the simplest form of cocycle. Perfect nonlinear (PN) 1D functions determine coboundaries with balanced partial derivatives. We use this property to define 2D PN and differentially k-uniform functions. We list the known PN permutations of GF(pa) as specific 2D PN coboundaries and show ∂ has an inverse for these PN functions. There are many more families of 2D PN cocycles on GF(pa) than those arising as coboundaries, even when p=2 (no 1D PN functions for p=2 exist; APN is the best possible). These ideas can be extended to include APN and differentially k-uniform 2D cocycles.
Published in:
Information Theory, 2002. Proceedings. 2002 IEEE International Symposium on
Date of Conference: 2002