DynaSubVAE: Adaptive Subgrouping for Scalable and Robust OOD Detection

TL;DR

DynaSubVAE is a novel adaptive framework that dynamically models latent subgroups in data, improving out-of-distribution detection especially for emerging and heterogeneous populations.

Contribution

It introduces a dynamic, non-parametric clustering-based approach within a VAE to adaptively detect OOD samples and model subpopulations in evolving data.

Findings

Achieves competitive near-OOD and far-OOD detection performance.

Excels in class-OOD scenarios with missing classes during training.

Outperforms standalone clustering methods like GMM and KMeans++.

Abstract

Real-world observational data often contain existing or emerging heterogeneous subpopulations that deviate from global patterns. The majority of models tend to overlook these underrepresented groups, leading to inaccurate or even harmful predictions. Existing solutions often rely on detecting these samples as Out-of-domain (OOD) rather than adapting the model to new emerging patterns. We introduce DynaSubVAE, a Dynamic Subgrouping Variational Autoencoder framework that jointly performs representation learning and adaptive OOD detection. Unlike conventional approaches, DynaSubVAE evolves with the data by dynamically updating its latent structure to capture new trends. It leverages a novel non-parametric clustering mechanism, inspired by Gaussian Mixture Models, to discover and model latent subgroups based on embedding similarity. Extensive experiments show that DynaSubVAE achieves competitive performance in both near-OOD and far-OOD detection, and excels in class-OOD scenarios where an entire class is missing during training. We further illustrate that our dynamic subgrouping mechanism outperforms standalone clustering methods such as GMM and KMeans++ in terms of both OOD accuracy and regret precision.