SupSiam: Non-contrastive Auxiliary Loss for Learning from Molecular Conformers
This work addresses the challenge of developing reliable E3NNs for molecular conformers in drug discovery, particularly for small-data programs, though it appears incremental as it builds on existing Siamese network and E3NN methods.
The paper tackled the problem of learning embeddings for molecular conformers using Siamese networks, finding that a non-contrastive auxiliary task improves manifold smoothness and aids supervised training of Euclidean neural networks (E3NNs) for drug-activity prediction tasks while maintaining performance metrics.
We investigate Siamese networks for learning related embeddings for augmented samples of molecular conformers. We find that a non-contrastive (positive-pair only) auxiliary task aids in supervised training of Euclidean neural networks (E3NNs) and increases manifold smoothness (MS) around point-cloud geometries. We demonstrate this property for multiple drug-activity prediction tasks while maintaining relevant performance metrics, and propose an extension of MS to probabilistic and regression settings. We provide an analysis of representation collapse, finding substantial effects of task-weighting, latent dimension, and regularization. We expect the presented protocol to aid in the development of reliable E3NNs from molecular conformers, even for small-data drug discovery programs.