News

Paulo Oliva has a new paper with Dr Andersson Silva and Prof. Ricardo Silva of the Federal University of Pernambuco in Brazil on "An approach using BRKGA for optimizing graph neural networkoptimizing graph neural network architectures". It is published in the Journal of Evolutionary Intelligence.

The paper was a collaboration that came about as a result of Prof. Ricardo Silva's earlier visit to QMUL. It provides a practical way to improve graph neural network (GNN) architectures by efficiently exploring the design space of possibilities. GNNs are neural networks that take as input and manipulate graphs (collections of nodes and the links between them) which are important data structures used to model many practical problems from improving transport networks to protein folding.

Abstract

Graph Neural Networks (GNNs) have become essential for learning from graph-structured data, achieving strong results across node classification, link prediction, and graph-level tasks. However, identifying effective GNN architectures remains difficult due to the heterogeneous and high-dimensional nature of their search spaces, which combine discrete architectural operators with continuous hyperparameters. This paper introduces BRKGA-GNN, a Neural Architecture Search framework that leverages the Biased Random-Key Genetic Algorithm (BRKGA) to efficiently explore mixed discrete–continuous GNN design spaces. The novelty of our approach lies in three components: (i) a continuous random-keys encoding
scheme, (ii) a parameterized decoder capable of translating random-key vectors into heterogeneous GNN topologies, and (iii) a rich and modular search space integrating aggregators, propagation mechanisms, normalization layers, activation functions, and pooling operators. BRKGA provides a principled and computationally efficient mechanism for navigating large architectural spaces, enabling smooth exploration through a continuous genotype while preserving structural validity through deterministic decoding. Experiments conducted on three benchmark citation networks (Cora, Citeseer, PubMed)
demonstrate that BRKGA-GNN achieves competitive or superior performance compared to representative NAS baselines, including evolutionary, reinforcement-learning, and weight-sharing methods, while maintaining low variance and stable convergence behavior. The results indicate that BRKGA-GNN constitutes a robust and scalable framework for automated GNN design, highlighting the effectiveness of combining random-keys encoding with a structured decoder for exploring complex architectural search spaces.

Reference

Silva, A.A., Silva, R.M.A. & Oliva, P. An approach using BRKGA for optimizing graph neural network architectures. Evol. Intel. 19, 100 (2026). https://doi.org/10.1007/s12065-026-01208-0