Riemann Primitives and Chrysalis

130 views
Skip to first unread message

Ian M

unread,
Jun 10, 2017, 5:02:29 PM6/10/17
to Cryptanalytic algorithms
I wanted to share current work with the group for review.  My colleague and I have been working to engineer a primitive to aid researchers in modeling post-quantum schemes using the same structure as a quantum computer.  We've chosen to call this a Riemann primitive as the 1:1 correspondence is between the Riemann sphere and Bloch sphere.  I've also chosen to term any cryptographic-secure points as Hollenbeck points, similar to that of Koblitz curves.


I've also attached it to this post.

Please keep in mind this is a working draft, and does not contain any formal proof or reduction, semantic security or otherwise.  This working draft does include propositions, definitions, and theorems I think will be necessary for a formally valid security reduction.  The overall parameters for implementing the Riemann primitive were left either vague, generalized, or not included.  This was done intentionally to help maintain a level of adaptability in using the primitive, though examples are included for lattice cryptography in general.

v/r,
Ian
Riemann Primitives and Chrysalis.pdf
Reply all
Reply to author
Forward
0 new messages