Molecular representation learning lays the foundation for drug discovery. However, existing methods suffer from poor out-of-distribution (OOD) generalization, particularly when data for training and testing originate from different environments.

For instance, patients' diagnosis varies substantially across medical centers but collaboratively form long-tailed distributions for

The code is available at Contrastive-Sampling .

In the field of natural language processing, the prevalent approach involves fine-tuning pretrained language models (PLMs) using local samples.

On one hand, it operates as a general data augmentation strategy.

Lijun Zhang is the corresponding author.

Gaussian mechanism borrowed from differential privacy.

While several previous works have focused on classifying close-set samples and detecting open-set samples during testing, it's still essential to be able to classify unknown subjects as human beings.