Searching Search Spaces: Meta-evolving a Geometric Encoding for Neural Networks
This work addresses a bottleneck in evolutionary algorithms for neural networks, offering an incremental improvement in encoding efficiency for researchers in neuroevolution.
The paper tackled the problem of inefficient neural network representations in evolutionary policy search by meta-evolving a distance function for the Geometric Encoding (GENE) method, resulting in improved performance over direct encoding and hand-crafted distances, with generalization to unseen problems.
In evolutionary policy search, neural networks are usually represented using a direct mapping: each gene encodes one network weight. Indirect encoding methods, where each gene can encode for multiple weights, shorten the genome to reduce the dimensions of the search space and better exploit permutations and symmetries. The Geometric Encoding for Neural network Evolution (GENE) introduced an indirect encoding where the weight of a connection is computed as the (pseudo-)distance between the two linked neurons, leading to a genome size growing linearly with the number of genes instead of quadratically in direct encoding. However GENE still relies on hand-crafted distance functions with no prior optimization. Here we show that better performing distance functions can be found for GENE using Cartesian Genetic Programming (CGP) in a meta-evolution approach, hence optimizing the encoding to create a search space that is easier to exploit. We show that GENE with a learned function can outperform both direct encoding and the hand-crafted distances, generalizing on unseen problems, and we study how the encoding impacts neural network properties.