Advances in Out-of-Distribution Detection and Hyperbolic Embeddings

The field of machine learning is moving towards developing more robust and reliable models, with a focus on out-of-distribution (OOD) detection and hyperbolic embeddings. Recent research has shown that hyperbolic geometry can be effectively used for OOD detection, allowing for more accurate identification of samples that do not belong to the training distribution. Additionally, there is a growing interest in developing methods that can dynamically adapt to new data and update their latent structure to capture emerging trends. Noteworthy papers include: DIsoN, which proposes a decentralized isolation network for OOD detection in medical imaging, and Balanced Hyperbolic Embeddings, which introduces a hyperbolic class embedding algorithm for OOD detection. DynaSubVAE is also notable for its dynamic subgrouping variational autoencoder framework that jointly performs representation learning and adaptive OOD detection.

Sources

Efficient Fireworks Algorithm Equipped with an Explosion Mechanism based on Student's T-distribution

Hyperbolic Dual Feature Augmentation for Open-Environment

DIsoN: Decentralized Isolation Networks for Out-of-Distribution Detection in Medical Imaging

Improving Out-of-Distribution Detection via Dynamic Covariance Calibration

Optimizing Latent Dimension Allocation in Hierarchical VAEs: Balancing Attenuation and Information Retention for OOD Detection

Balanced Hyperbolic Embeddings Are Natural Out-of-Distribution Detectors

DynaSubVAE: Adaptive Subgrouping for Scalable and Robust OOD Detection

Continual Hyperbolic Learning of Instances and Classes

Built with on top of