Structural Plasticity as Active Inference: A Biologically-Inspired Architecture for Homeostatic Control

TL;DR

This paper presents SAPIN, a biologically-inspired neural architecture that combines local prediction error minimization with structural plasticity, enabling effective reinforcement learning on the Cart Pole task.

Contribution

Introduces SAPIN, a novel model integrating active inference principles with morphological plasticity for homeostatic control in neural networks.

Findings

Successfully solves the Cart Pole task with stable policies.

Structural plasticity enhances learning adaptability.

Locked parameters maintain high success rates over episodes.

Abstract

Traditional neural networks, while powerful, rely on biologically implausible learning mechanisms such as global backpropagation. This paper introduces the Structurally Adaptive Predictive Inference Network (SAPIN), a novel computational model inspired by the principles of active inference and the morphological plasticity observed in biological neural cultures. SAPIN operates on a 2D grid where processing units, or cells, learn by minimizing local prediction errors. The model features two primary, concurrent learning mechanisms: a local, Hebbian-like synaptic plasticity rule based on the temporal difference between a cell's actual activation and its learned expectation, and a structural plasticity mechanism where cells physically migrate across the grid to optimize their information-receptive fields. This dual approach allows the network to learn both how to process information (synaptic weights) and also where to position its computational resources (network topology). We validated the SAPIN model on the classic Cart Pole reinforcement learning benchmark. Our results demonstrate that the architecture can successfully solve the CartPole task, achieving robust performance. The network's intrinsic drive to minimize prediction error and maintain homeostasis was sufficient to discover a stable balancing policy. We also found that while continual learning led to instability, locking the network's parameters after achieving success resulted in a stable policy. When evaluated for 100 episodes post-locking (repeated over 100 successful agents), the locked networks maintained an average 82% success rate.