Averaging acyclic key-chain hybrids #
The signed gap of a key-chain experiment telescopes over links, then over secret-key bits. Both coordinates are averaged over fixed dyadic ranges. Unused coordinates reject in both worlds, including zero links and empty secret words. No parameter-dependent best coordinate is selected.
These are exact game identities. Connecting the averaged game to an actual native IND-CPA adversary is a separate reduction obligation.
Main definitions #
Cslib.Crypto.Homomorphic.Scheme.keyChainHybridGame: the two-coordinate hybrid average.
Main results #
Cslib.Crypto.Homomorphic.Scheme.publicKeyLinkHybrid_gap: signed bitwise telescoping.Cslib.Crypto.Homomorphic.Scheme.publicKeyChainHybrid_gap: the full signed two-coordinate identity with its explicit dyadic loss.
References #
Telescope the bit hybrids of one link. The cap can be any upper bound on the common supported secret length; it need not equal that length.
Average reverse link indices and forward bit indices. Inactive coordinates return the same constant in both worlds. This is a distributional definition, not an unchecked sampler.
Instances For
The real-to-zero public-key gap is the signed gap of one uniformly selected link/bit hybrid, multiplied by the two dyadic range sizes.