Maya-Morphe P4: Dharana — EFC Trajectory Dynamics in Morphogenetic Neural Networks

Birla Institute of Technology and Science, Pilani

Indexed indatacite

Abstract

Maya-Morphe P4: Dharana characterises the full EFC (Edge-Field Concordance) trajectory dynamics in morphogenetic neural networks — the first epoch-by-epoch measurement of EFC over 250 training epochs across 5 scales (86,931 to 266,071 cells). Key findings:- EFC follows a universal U-shaped trajectory: exploratory deepening (E1-175, trough -0.13 to -0.19), temperature-driven recovery (E175-225), convergence to EFC=0.000 at T=0.1 (E225-250)- FRR=100% confirmed scale-invariant from 86k to 266k cells. FIXED_TOPOLOGY baseline: 0%- 50k scale shows anomalously shallow EFC trough (-0.1294), consistent with Manicka & Levin (2025) optimal field sensitivity range- Gumbel-Softmax temperature identified as the universal…

Citation impact

42
total citations
FWCI
Percentile
References
21
Too recent for citation history.

Authors

1

Topics & keywords

Keywords
  • Convergence (economics)
  • Trajectory
  • Artificial neural network
  • Scale (ratio)
  • Dynamics (music)
  • Sensitivity (control systems)
No related works found for this paper.