Lin Hsin Hsin Quantum Security Center
Post-Quantum Cryptography
There are three standardized PQC algorithms
| Standard |
Former name
As if a name change improves mathematics & algorithm |
Type |
Mathematics |
| ML-KEM (FIPS 203) |
CRYSTALS-Kyber |
Key encapsulation |
Module-LWE |
| ML-DSA (FIPS 204) |
CRYSTALS-Dilithium |
Digital signatures |
Module-LWE + Module-SIS |
| SLH-DSA (FIPS 205) |
SPHINCS+ |
Digital signatures |
Hash-based (not lattice) |
WHY LATTICE-BASED CRYPTOGRAPHY FAILS
Lattice-based schemes are built on the hardness of
🤡
LWE
🤡 MLWE
🤡 Module-SIS
Lattice-based algorithm hardness is the foundational security,
assumption that specific geometric problems in high-dimensional grids (lattices) are computationally intractable for both classical and quantum computers
The lattice geometry admits a classical, device-agnostic shortcut that bypasses the hardness assumption entirely.
The reduction is not an attack on the lattice — it is a proof that the security proof's reduction step is invalid
⚠ No dimension increase fixes this
⚠ No module rank change fixes this
⚠ The proof is the vulnerability
It is:
⚠ Nothing
⚠ Not for parameter selection
⚠ Not for security justification
⚠ Not for anything that touches a deployed system
It's cited in the security discussion of
FIPS 203 as a "structural argument" — meaning: "the problem class has no known weak instances." That's it. The actual security number comes from the lattice estimator against BKZ, which has
nothing to do with the reduction
Every post-quantum standard was chosen because the classical alternatives were broken.
They are not safe
They are the same category of break
expressed in different mathematics
Hash-based: structural collision
Lattice-based: invalid reduction
The entire PQC pipeline
NIST, ETSI, ISO
is built on assumptions my theorems invalidate