Exploration of Novel Neuromorphic Methodologies for Materials Applications

TL;DR

This paper evaluates neuromorphic computing strategies, reservoir and hyperdimensional computing, for materials property prediction using graph neural network tasks, demonstrating hyperdimensional computing's effectiveness in representing molecular graphs.

Contribution

It introduces the application of neuromorphic approaches to materials science graph tasks, comparing reservoir and hyperdimensional computing for the first time in this context.

Findings

Hyperdimensional computing improves molecular graph representation.

Neuromorphic strategies address GNN limitations like over-smoothing.

Results show effective performance in bandgap classification and regression.

Abstract

Many of today's most interesting questions involve understanding and interpreting complex relationships within graph-based structures. For instance, in materials science, predicting material properties often relies on analyzing the intricate network of atomic interactions. Graph neural networks (GNNs) have emerged as a popular approach for these tasks; however, they suffer from limitations such as inefficient hardware utilization and over-smoothing. Recent advancements in neuromorphic computing offer promising solutions to these challenges. In this work, we evaluate two such neuromorphic strategies known as reservoir computing and hyperdimensional computing. We compare the performance of both approaches for bandgap classification and regression using a subset of the Materials Project dataset. Our results indicate recent advances in hyperdimensional computing can be applied effectively to better represent molecular graphs.