No out-of-band channel: StOMR sets up shared state for OMR and shrinks sender signals in anonymous comms.

Debate still fails in the ROM: this paper proves ZK AI oversight impossible there, then salvages it with signed oracles and CRHs.

Even expander codes can hit GV, with certified finite-length bounds, failure probabilities, and no decoding fairy dust.

Hand-coded AArch64 NEON for batched LSH, with a 4-way SPHINCS+ core that speeds signing without changing signatures.

Can KOALA really buy public-coin differing-input obfuscation, or is this just witness-encryption laundering with extra steps?

What if the code hull isn’t trivial? This extends Ducas-Gibbons to Construction A lattices and still runs in 2^{O(n)}.

Bootstrapping gets fused into one GPU megakernel here, because apparently TFHE needed fewer global memory trips.

Post-quantum X3DH that survives reverse firewalls and kills subliminal channels, or just another isogeny handshake?

Finds a downgrade hole in ITU-T X.509 Catalyst PQC migration, where a classical CA key can swap in classical-only certs.

Preprint on a subexponential distinguisher for binary LWE with only n/log log n samples, because linear samples were too mainstream.

CAROUSEL paper: FHE private vector search that scores the whole corpus, sketches encrypted scores, and farms it out to GPUs.

Explicit Humbert-discriminant-36 equations, plus the part where level-6 makes splitting detection less pleasant than the usual folklore.

TFHE bootstrapping gets an offline lattice search plus LUT shifting, so the online path stays boring while noise drops.

Grinding challenges on a GPU, then piping FAEST v3 through CPU and GPU, because apparently signing needed a sidecar.

MiniSHAKE defines Keccak-based hash/XOFs on Keccak-p[800,12], for when 1600 bits was apparently too much.

Can X-Wing fit on Cortex-M85, and what do the cycle counts, storage, and timing leaks actually look like?

Where in the net does secret-bearing leakage actually appear, and does GlimpseDNN finally show it layer by layer?

FFT multiplication finally wins earlier on ARM, with tighter Barrett bounds and an OpenSSL RSA speedup to prove it.

Paper on parallel radix digit decomposition and sign evaluation for CKKS, with log-depth carry fix and fewer moduli, alas.

First masked Knuth-Yao Gaussian sampler, plus a generic finite-precision compiler and Cortex-M4 numbers for lattice schemes.

Multiparty HSS from two-party semi-private HSS, via party virtualization and DCR/class-group assumptions.

Output-adaptive ABE for circuits from LWE, the first of its kind without iO or witness encryption, because selective security was too neat.

Can you get fast large-field linear codes with just adds and tiny coefficients? This paper says mostly yes, with distance guarantees.

Power analysis breaks masked Falcon via share recombination leakage, with chosen-message and passive variants.

Turns Shamir into BIP39 mnemonics, with checksum/MAC validation and paper-only recovery, instead of another binary blob.

Paper on ATLAS, a compact module-MLWR signature scheme with AVX2 and SHAKE/SM3 backends, because more lattice acronyms were needed.

Keygen cost isn't the lattice, it's rejection and secret lifetimes on Cortex-M4, per a controlled study of LWR, ML-KEM, HAETAE.

Separates Fiat-Shamir PoW into grinding vs diagonalization, then shows one PoW can’t do both when verifiers are cheaper.

eBPF-based Linux crypto inventory, it uses randomness as a timing anchor to spot keygen and PQC ops at runtime.

ePrint on PQC OTAR over CCSDS links, showing endpoint processing time changes retransmission paths and timing.