Self-Motivated Growing Neural Network for Adaptive Architecture via Local Structural Plasticity

TL;DR

This paper presents SMGrNN, a neural network that dynamically adapts its architecture during learning through local signals, improving control performance and stability without manual tuning.

Contribution

Introduction of a local structural plasticity module enabling online topology growth and pruning in neural networks for reinforcement learning.

Findings

Achieves comparable or better rewards than fixed architectures.

Reduces network size while maintaining performance.

Enhances reward stability through adaptive topology.

Abstract

Control policies in deep reinforcement learning are often implemented with fixed-capacity multilayer perceptrons trained by backpropagation, which lack structural plasticity and depend on global error signals. This paper introduces the Self-Motivated Growing Neural Network (SMGrNN), a controller whose topology evolves online through a local Structural Plasticity Module (SPM). The SPM monitors neuron activations and edge-wise weight update statistics over short temporal windows and uses these signals to trigger neuron insertion and pruning, while synaptic weights are updated by a standard gradient-based optimizer. This allows network capacity to be regulated during learning without manual architectural tuning. SMGrNN is evaluated on control benchmarks via policy distillation. Compared with multilayer perceptron baselines, it achieves similar or higher returns, lower variance, and task-appropriate network sizes. Ablation studies with growth disabled and growth-only variants isolate the role of structural plasticity, showing that adaptive topology improves reward stability. The local and modular design of SPM enables future integration of a Hebbian plasticity module and spike-timing-dependent plasticity, so that SMGrNN can support both artificial and spiking neural implementations driven by local rules.