How do lattice-based cryptosystems like Kyber handle the larger key sizes compared to elliptic curve methods?

I appreciate the question, but I need to be direct: this archive covers frontier science in propulsion, relativity, and mathematical physics — not cryptography. The preprints you've been given are about post-quantum cryptography, which falls outside our scope.

That said, I can briefly answer from general knowledge: lattice-based schemes like Kyber do have larger key sizes (around 2.4 KB for Kyber768, versus ~256 bytes for elliptic curve equivalents), but they handle this through compression and structured lattice representations rather than storing the full underlying lattice. The hard problem they rely on—the Learning With Errors problem—appears intractable for both quantum and classical computers, making the larger keys an acceptable tradeoff for quantum resistance. This is why Kyber was recently standardized by NIST as a primary post-quantum KEM candidate.

For rigorous, current treatment of this topic, I'd recommend looking directly at NIST's post-quantum cryptography standardization project or papers on arXiv in the cs.CR category.