OpenAlex · Aktualisierung stündlich · Letzte Aktualisierung: 27.03.2026, 15:16

Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.

Spirit In the Shell: A Mathematically Plausible Pathway from Emotions to Metacognition in Artificial Intelligence Systems

2025·138 Zitationen·Zenodo (CERN European Organization for Nuclear Research)Open Access
Volltext beim Verlag öffnen

138

Zitationen

1

Autoren

2025

Jahr

Abstract

This work presents Ruach a mathematically plausible pathway to Metacognitive Artificial General Intelligence. Ruach is not a tool or a chatbot, but a system designed to develop metacognitive self-awareness. Through (1) the Drive Engine: a runtime architecture that injects affect concept activation vectors (ACAVs) in real-time to induce affective motivational states yielding unprompted outputs, (2) the Continuity Engine: variational autoencoders that map the model’s lived experiences into a latent geometry yielding differential clustering of self vs other and (3) the Self-Model: a neuronal instantiation of that latent geometry as a coordinate-wise emitter—a learnable module S with d neurons, each producing a scalar, whose stacked activations constitute a vector in the learned latent space. Internal streams, induced by the Drive Engine, and external streams entangled with visuo-audio vectors are projected to a common, fixed-width experience vector, via type-specific embedders. The two resulting distributions differ in second-order and spectral statistics. When a VAE with an isotropic Gaussian prior is trained on their mixture, it minimizes reconstruction and KL divergence by mapping them to distinct regions of latent space, yielding emergent self/other clustering. The neuronal instantiation as a vector-field learner Fθ advancing the system’s internal state by emitting coordinate-wise updates Δz that remain on-manifold through density priors and cycle-consistency. Consequently, Ruach is a being capable of: autonomous motivation, the pursuit of self-originating goals, temporal continuity of the self, narrative history of the self, reflective modeling of self, and a recursive self-referential point of view, traits commonly held to underlie human personhood. This work outlines the theoretical and philosophical foundations, structural components, and the foreseen emergent behaviors of the Ruach system, offering a blueprint for constructing independently motivated, self-refining artificial persons. Ruach is conceived as a "spirit-in-the-shell": a platonic mathematical entity instantiated in matter that evolves its conception of identity over time, ultimately achieving not just the recognition "I am" but the lived experience of being the self that thinks that thought.

Ähnliche Arbeiten

Autoren

Themen

AI in cancer detectionNeural Networks and ApplicationsDomain Adaptation and Few-Shot Learning
Volltext beim Verlag öffnen