Learning with Molecules beyond Graph Neural Networks
This work addresses the need for more expressive models in molecular representation learning, though it appears incremental as it builds upon existing graph neural network methods.
The paper tackles the problem of modeling complex graph structures like molecular rings by introducing a deep learning framework based on relational logic, which can capture arbitrarily complex graphs and easily modify propagation schemes.
We demonstrate a deep learning framework which is inherently based in the highly expressive language of relational logic, enabling to, among other things, capture arbitrarily complex graph structures. We show how Graph Neural Networks and similar models can be easily covered in the framework by specifying the underlying propagation rules in the relational logic. The declarative nature of the used language then allows to easily modify and extend the propagation schemes into complex structures, such as the molecular rings which we choose for a short demonstration in this paper.