Draft — This content has not been approved for publication.

← Back to Publications

ActPC-Geom: Towards Scalable Online Neural-Symbolic Learning via Accelerating Active Predictive Coding with Information Geometry

Author: Ben Goertzel
Year: 2025
Venue: arXiv:2501.04832
Links: arXiv abstract

Summary

A more geometric and scaling-oriented companion to the discrete ActPC work. It explores how active predictive coding can be accelerated using information geometry, pushing the predictive-coding line toward online neural-symbolic learning that is more plausible as a reusable subsystem inside a larger AGI architecture.

Relevance to Hyperon

This paper strengthens the predictive-coding side of PRIMUS Full and complements ActPC-Chem by shifting from discrete algorithmic chemistry to scalable geometric learning dynamics.

Key References



Discussion