PHILIA-DAN and TAN: A Measurement Study of Pairwise and Multi-Source Coupling in Recursive Artificial Neurons
(Description) This paper presents PHILIA-DAN (Dual Artificial Neuron) and PHILIA-TAN (Triple Artificial Neuron), minimal toy models for studying pairwise and multi-source coupling in a recursive artificial neuron system (PHILIA v9). Building on single-neuron passive stability demonstrated in PHILIA-SAN, this study measures coupling effects across two-neuron (DAN) and three-neuron (TAN) configurations using real datasets from CERN dielectron (99,915 events), ATLAS Higgs (693,636 events), and NYC Taxi Trip Duration (1,450,031 events after filtering). Key findings: SR→Φ coupling is inactive due to S saturation. SR→C coupling produces measurable ignition acceleration with a linear response zone at ε=0.001–0.010. TAN extends DAN to pairwise additive multi-source coupling without introducing new dynamics. Control experiments (shuffled sequence, pure additive coupling) confirm that observed effects are conditioned on input sequence order and the (1−S) damping mechanism. All results are from local hardware execution (Intel N100). Description, not proof. — PHILIA OS
Authors
- 두섭 신
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-04-01
- DOI
- https://doi.org/10.5281/zenodo.19363532
- Primary Topic
- stochastic dynamics and bifurcation
- Type
- article
- Field-Weighted Citation Impact
- 0.00