arXiv cs.AI endorsement request — Independent researcher, single-author submission

Dear arXiv endorsers,

I am an independent researcher based in Seoul, Republic of Korea, preparing my first arXiv submission (cs.AI primary, q-bio.NC cross-list) as a single author. I would be grateful for an endorsement.

Paper title

Sentient Topology: Plasticity-Aware Affective Representation in a Humanities-Grounded Sensory Associative Network

Summary

The paper models artificial emotion as the topological shape of activation in an 8,000-node Sensory Associative Network (SAN) built from a 200-work Project Gutenberg corpus (~22.3M tokens). A five-dimensional topological signature (Density, Symmetry, Centrality, Depth, Boundary) is extracted with a pure-Python Z_2 persistent-homology solver; a six-mechanism plasticity engine reproduces the Thompson-Spencer habituation curve, the Walker-Skowronski Fading Affect Bias, and surprise sensitization on the same SAN. The empirical core is an intra-corpus affect-trajectory differentiation protocol: parameterizing SAN propagation with Panksepp’s six primary affects (LUST excluded) plus a uniform baseline, we run 7 x 6 = 42 deterministic propagations across six stimuli. The Boundary axis emerges as the primary observable (13 of 21 affect pairs systematically differ); a continuum from constriction (RAGE/FEAR/PANIC) through baseline to expansion (PLAY/SEEKING/CARE) emerges on mean Boundary as a suggestive analog to Panksepp’s approach/avoidance axis. A +/-50% sensitivity analysis identifies two brittle dominant channels. The framework makes no claim of conscious felt experience and no claim to validate Panksepp’s framework; we report what a committed affect-to-SAN parameter mapping produces.

Author

• Independent Researcher, Seoul, Republic of Korea

• ORCID: 0009-0002-3258-2466

• Code & data: github.com/HocH88/sentient-topology

• Single author, no institutional affiliation

The GitHub repository contains the full LaTeX source (paper/main.tex), both generated figures, the 8,000-node SAN JSON, the corpus manifest, and the Python pipeline for all four affect-trajectory stages plus the plasticity experiments.

What I would appreciate

• An endorsement for the cs.AI subject class.

• If you are unable but know someone who would be appropriate, any pointer is most welcome.

Thank you for your time.

Best regards, Hochul HAN

email: hochhan88@gmail.com