Classification via score-based generative modelling
This work addresses classification challenges, particularly in high dimensions and imbalanced data, by applying a known method (score-based modeling) to enhance performance, making it incremental.
The paper tackled classification by using score-based generative modeling to characterize data distributions and generate samples, achieving improved binary classification performance and robustness in high-dimensional and imbalanced scenarios.
In this work, we investigated the application of score-based gradient learning in discriminative and generative classification settings. Score function can be used to characterize data distribution as an alternative to density. It can be efficiently learned via score matching, and used to flexibly generate credible samples to enhance discriminative classification quality, to recover density and to build generative classifiers. We analysed the decision theories involving score-based representations, and performed experiments on simulated and real-world datasets, demonstrating its effectiveness in achieving and improving binary classification performance, and robustness to perturbations, particularly in high dimensions and imbalanced situations.