Public code artifacts from research work.
Implementations and supporting code for specific research papers.
| Anchor | Description |
|---|---|
| gupta2026architecture | Architecture-induced recoverability bias in differentiable symbolic regression (arXiv:2604.23256, MLSP 2026) |
| gupta2026saturation | The saturation number is not bounded by the harmonic index (arXiv:2606.15761) — refutation of TxGraffiti Conjecture 4 (arXiv:2507.17780) |
| Anchor | Description |
|---|---|
| sanadhya2008 | 22- and 24-step SHA-256 collision attacks |
| mendel2011 | 27/32-step SHA-256 collisions via SAT + novel findings (32K collision family, 10× phase-seed speedup, W[7] characterization) |
| mendel2013improving | 28/31/38-step SHA-256 collisions via SAT on dense characteristics |
| eichlseder2014 | 38-step SHA-512 SFS collision via SAT + GnD plateau analysis |
| lamberger2011 | 46-step SHA-256 second-order differential: message-modification analysis (two-layer decomposition, revised 2⁴⁷·⁶ complexity) |
| de2006finding-workfactor | SHA-1 work-factor calculator (De Cannière–Rechberger 2006 §III-C), first independent reimplementation |
| gupta2026rank | Rank ceiling for twiddle-perturbation faults on the forward NTT of ML-KEM and ML-DSA (ePrint 2026/1188) |
| zhang2026-verification | SFS collision verification + corrected Table 8 characteristic for Zhang–Li–Gao–Wang, EUROCRYPT 2026 |
Stand-alone utilities not tied to a single paper. Organized by domain.
| Tool | Description |
|---|---|
| qbudget | Report T-count and AND-uncomputation accounting for compute-uncompute Grover oracles in Qiskit or Q# |