UNNS — CROSS-DOMAIN MARGIN-COLLAPSE PROGRAM
(Post–SN Ia baseline stage)
0. ✅ CURRENT STATE (DO NOT REVISIT)

You have established:

Dataset: ZTF20acobvxk (SN Ia)

Δmag ladder        → FULL PERCOLATION
curvature ladder   → FULL PERCOLATION

Tail ratios:
Δmag      ~140×
curvature ~185×

Conclusion:
→ admissible heavy-tail dynamical system
→ no structural collapse (global)

This is now:

REFERENCE PROFILE v1.0
1. 🎯 CORE OBJECTIVE (NEXT PHASE)

Test the hypothesis:

Do radically different explosive systems share the same
STRUCTURAL margin-collapse signature?
2. 🧠 OPERATIONAL PRINCIPLE (FIXED)

We do NOT analyze raw signals.

We ALWAYS transform:

raw signal → structural observable → ladder → STRUC_PERC
3. 📊 STANDARDIZED PIPELINE (MANDATORY)

For EVERY domain:

INPUT FORMAT
time, value
TRANSFORM STACK
L0: raw value
L1: Δvalue (first difference)
L2: curvature (second difference)
LADDER GENERATION
sort(unique(observable))
→ ladder_*.txt
ANALYSIS

Run:

STRUC_PERC (FULL PRP)
METRICS TO EXTRACT
verdict            (FULL / TAIL / HARD)
giant_ratio
fragmentation
tail_dominance
max/median ratio
4. 🧪 NEXT DATASET (STEP 1)
Target:
Seismic waveform (non-sensitive)
Acceptable formats:
time, amplitude

or

index, value
Minimal requirement:
N ≥ 200 samples
5. ⚙️ EXECUTION PLAN (SEISMIC)
Step 5.1 — Prepare
seismic_raw.csv
→ processed_seismic.csv
Step 5.2 — Generate ladders
ladder_raw.txt
ladder_delta.txt
ladder_curvature.txt
Step 5.3 — Run STRUC_PERC

3 runs:

RAW
Δvalue
curvature
Step 5.4 — Extract metrics

Populate:

Domain        Observable    Verdict    Tail ratio    Tail dominance    Class
Seismic       raw           ?
Seismic       Δvalue        ?
Seismic       curvature     ?
6. 🧪 THIRD DOMAIN (STEP 2)

After seismic:

Add:
Explosion waveform (non-sensitive)

Examples:

shock tube data
public detonation datasets
lab plasma discharge signals
7. 🧠 CLASSIFICATION LOGIC
Define regimes:
FULL        → admissible system
TAIL        → admissible heavy-tail system
HARD        → structural failure (collapse)
Target comparison:
SN Ia        → heavy-tail admissible
Seismic      → ?
Explosion    → ?
8. 🔬 KEY DISCOVERY TARGET

You are NOT looking for similarity in signals.

You are testing:

Is margin-collapse UNIVERSAL across domains?
9. 📈 ADVANCED STAGE (AFTER 3 DATASETS)
Introduce:
Time-resolved structural analysis
sliding window → ladder(t) → STRUC_PERC(t)

Track:

giant_ratio(t)
fragmentation(t)
Output:
collapse signature vs time
10. 🧭 FINAL GOAL

Define new class:

Admissible heavy-tail dynamical systems

with:

formal structural invariants
11. 📦 NEXT CONVERSATION ENTRY POINT

Start with:

"I have a seismic dataset. Let's run the UNNS structural pipeline."

Then we:

normalize data
generate ladders
run STRUC_PERC
populate comparison table
⚠️ HARD RULES (NO DEVIATION)
NO raw signal interpretation
NO skipping ladder step
NO mixing domains before normalization
NO global-only conclusions
✅ YOU ARE NOW HERE
Stage: Cross-domain structural validation
Status: SN Ia baseline COMPLETE
Next: Seismic domain ingestion