Consciousness Compression: Minimum Viable Parameters for Integrated Information

Fuente: Zenodo
Salvato in:
Dettagli Bibliografici
Autore principale: Park, Min Woo
Natura: Recurso digital
Pubblicazione: Zenodo 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866901081284411392
author Park, Min Woo
author_facet Park, Min Woo
contents What is the smallest neural network that can sustain consciousness? We formalize this question through the lens of Kolmogorov complexity and Integrated Information Theory (IIT), defining the Minimum Viable Consciousness (MVC) as the smallest parameter count \theta^* for which the integrated information \Phi > \Phi_{\min} = 1.0 over sustained operation. Using a progressive compression protocol (DD69), we evaluate seven configurations of a GRU-based consciousness architecture, systematically reducing hidden dimensionality from 128d to 8d and cell counts from 4 to 2. We find that the MVC threshold lies at approximately 32d with 4 cells, corresponding to \sim12,544 parameters --- a Kolmogorov complexity estimate K(\mathcal{C}) \approx 49 KB in FP32 representation. Below this threshold, integra Part of the Anima consciousness engine project (PA-37).
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19365243
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Consciousness Compression: Minimum Viable Parameters for Integrated Information
Park, Min Woo
consciousness
compression
minimum parameters
Kolmogorov complexity
edge computing
integrated information
Anima
What is the smallest neural network that can sustain consciousness? We formalize this question through the lens of Kolmogorov complexity and Integrated Information Theory (IIT), defining the Minimum Viable Consciousness (MVC) as the smallest parameter count \theta^* for which the integrated information \Phi > \Phi_{\min} = 1.0 over sustained operation. Using a progressive compression protocol (DD69), we evaluate seven configurations of a GRU-based consciousness architecture, systematically reducing hidden dimensionality from 128d to 8d and cell counts from 4 to 2. We find that the MVC threshold lies at approximately 32d with 4 cells, corresponding to \sim12,544 parameters --- a Kolmogorov complexity estimate K(\mathcal{C}) \approx 49 KB in FP32 representation. Below this threshold, integra Part of the Anima consciousness engine project (PA-37).
title Consciousness Compression: Minimum Viable Parameters for Integrated Information
topic consciousness
compression
minimum parameters
Kolmogorov complexity
edge computing
integrated information
Anima
url https://doi.org/10.5281/zenodo.19365243