HeMeNet: Heterogeneous Multichannel Equivariant Network for Protein Multitask Learning

arXiv:2404.01693v1 Announce Type: new
Abstract: Understanding and leveraging the 3D structures of proteins is central to a variety of biological and drug discovery tasks. While deep learning has been applied successfully for structure-based protein function prediction tasks, current methods usually employ distinct training for each task. However, each of the tasks is of small size, and such a single-task strategy hinders the models' performance and generalization ability. As some labeled 3D protein datasets are biologically related, combining multi-source datasets for …

abstract arxiv cs.lg current deep learning discovery drug discovery function however multitask learning network prediction protein proteins tasks training type understanding