Evolution imposes an inductive bias that alters and accelerates learning dynamics

TL;DR

This paper demonstrates that evolutionary processes create an inductive bias in neural networks, leading to altered and faster learning dynamics, and enabling rapid fine-tuning similar to innate biological structures.

Contribution

It introduces a method combining natural selection algorithms with online learning to evolutionarily condition neural networks, revealing evolution's role as an inductive bias for rapid learning.

Findings

Evolutionarily conditioned networks show unique learning dynamics.

Such networks can be rapidly fine-tuned to optimal performance.

Evolution acts as an inductive bias that accelerates learning.

Abstract

The learning dynamics of biological brains and artificial neural networks are of interest to both neuroscience and machine learning. A key difference between them is that neural networks are often trained from a randomly initialized state whereas each brain is the product of generations of evolutionary optimization, yielding innate structures that enable few-shot learning and inbuilt reflexes. Artificial neural networks, by contrast, require non-ethological quantities of training data to attain comparable performance. To investigate the effect of evolutionary optimization on the learning dynamics of neural networks, we combined algorithms simulating natural selection and online learning to produce a method for evolutionarily conditioning artificial neural networks, and applied it to both reinforcement and supervised learning contexts. We found the evolutionary conditioning algorithm, by itself, performs comparably to an unoptimized baseline. However, evolutionarily conditioned networks show signs of unique and latent learning dynamics, and can be rapidly fine-tuned to optimal performance. These results suggest evolution constitutes an inductive bias that tunes neural systems to enable rapid learning.