Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Guiding Genetic Programming with Graph Neural Networks

Piotr Wyrwiński, Krzysztof Krawiec

TL;DR

EvoNUDGE is proposed, which uses a graph neural network to elicit additional knowledge from symbolic regression problems to significantly outperform multiple baselines, including the conventional tree-based genetic programming and the purely neural variant of the method.

Abstract

In evolutionary computation, it is commonly assumed that a search algorithm acquires knowledge about a problem instance by sampling solutions from the search space and evaluating them with a fitness function. This is necessarily inefficient because fitness reveals very little about solutions -- yet they contain more information that can be potentially exploited. To address this observation in genetic programming, we propose EvoNUDGE, which uses a graph neural network to elicit additional knowledge from symbolic regression problems. The network is queried on the problem before an evolutionary run to produce a library of subprograms, which is subsequently used to seed the initial population and bias the actions of search operators. In an extensive experiment on a large number of problem instances, EvoNUDGE is shown to significantly outperform multiple baselines, including the conventional tree-based genetic programming and the purely neural variant of the method.

Guiding Genetic Programming with Graph Neural Networks

TL;DR

EvoNUDGE is proposed, which uses a graph neural network to elicit additional knowledge from symbolic regression problems to significantly outperform multiple baselines, including the conventional tree-based genetic programming and the purely neural variant of the method.

Abstract

In evolutionary computation, it is commonly assumed that a search algorithm acquires knowledge about a problem instance by sampling solutions from the search space and evaluating them with a fitness function. This is necessarily inefficient because fitness reveals very little about solutions -- yet they contain more information that can be potentially exploited. To address this observation in genetic programming, we propose EvoNUDGE, which uses a graph neural network to elicit additional knowledge from symbolic regression problems. The network is queried on the problem before an evolutionary run to produce a library of subprograms, which is subsequently used to seed the initial population and bias the actions of search operators. In an extensive experiment on a large number of problem instances, EvoNUDGE is shown to significantly outperform multiple baselines, including the conventional tree-based genetic programming and the purely neural variant of the method.

Paper Structure

This paper contains 14 sections, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Problem formulation
  3. The proposed method
  4. Motivations
  5. NUDGE
  6. The EvoNUDGE hybrid
  7. Related work
  8. Experiments
  9. The sample of SR problems
  10. GNN training
  11. Configurations of compared methods
  12. Results
  13. Performance on other benchmarks
  14. Discussion and Conclusions

Figures (2)

  • Figure 1: An example of a graph constructed by NUDGE.
  • Figure 2: The architecture of the GNN used for saliency estimation in NUDGE; $n$: the number of examples in the dataset that specifies the SR problem instance; $N$: the number of message passing iterations and of the GAT layers of the model.