Commit graph

1 commit

Author SHA1 Message Date
openresearch
02c815de8d docs(research): BraidStorm × TreeBraid × COUCH chiral batch pipeline
Connects three existing SilverSight components:
1. BraidStorm (BraidEigensolid.lean) — 8-strand braid, Sidon labels,
   chiral crossings σ_i^±1 → 2^8 = 256 configurations per run
2. TreeBraid — tree-organized braid, factorizes via σ_i σ_j = σ_j σ_i
   (|i-j|≥2), reduces 256 to ~64-128 unique configs
3. COUCH (GCCL.lean couchStable gate) — moving sofa constraint,
   geometric pre-filter (cheap, O(1) per config)

Pipeline: BraidStorm generates → TreeBraid factorizes →
COUCH filters geometrically → Sidon filters algebraically (dual
quaternion products, no tolerance band).

COUCH is the CHEAP filter (geometric). Sidon is the EXPENSIVE filter
(algebraic O(n²)). Running COUCH first rejects ~50% of configs,
halving the Sidon workload.

Final output: ~10-20 structurally meaningful configs per run
(from 256 raw). These are where the octagon principle could detect
the sofa's chromatic structure from the spectrum.

Hutter prize lesson: the batch doesn't COMPRESS 256→1 (conservation
law blocks that). It FILTERS 256→10-20 that are both geometrically
valid and structurally meaningful.

Also adds CHIRAL_BATCH_ENCODING.md (the general framework).
2026-07-04 19:38:41 +00:00