Evolving Self-Assembling Neural Networks: From Spontaneous Activity to Experience-Dependent Learning
This addresses the problem of adaptability in changing environments for AI systems, though it builds incrementally on prior Neural Developmental Programs.
The paper tackles the brittleness of static artificial neural networks by introducing Lifelong Neural Developmental Program (LNDP), a self-organizing model with synaptic and structural plasticity, and demonstrates its ability to learn from experiences in control tasks and adapt quickly in non-stationary environments.
Biological neural networks are characterized by their high degree of plasticity, a core property that enables the remarkable adaptability of natural organisms. Importantly, this ability affects both the synaptic strength and the topology of the nervous systems. Artificial neural networks, on the other hand, have been mainly designed as static, fully connected structures that can be notoriously brittle in the face of changing environments and novel inputs. Building on previous works on Neural Developmental Programs (NDPs), we propose a class of self-organizing neural networks capable of synaptic and structural plasticity in an activity and reward-dependent manner which we call Lifelong Neural Developmental Program (LNDP). We present an instance of such a network built on the graph transformer architecture and propose a mechanism for pre-experience plasticity based on the spontaneous activity of sensory neurons. Our results demonstrate the ability of the model to learn from experiences in different control tasks starting from randomly connected or empty networks. We further show that structural plasticity is advantageous in environments necessitating fast adaptation or with non-stationary rewards.