DP-BOA: Dirichlet-Process Birth-or-Assign for On-the-Fly Category Discovery

2026-07-15

AI Summary

This work addresses the problem of explicitly determining, in an online setting, whether each incoming sample should be assigned to an existing category or used to instantiate a novel one. To this end, the authors propose a novel approach based on an online Dirichlet process Gaussian mixture model, which for the first time treats the emergence of a new category as an explicit decision option grounded in independent statistical evidence, rather than as a fallback mechanism triggered by matching failure. The method employs a Normal-Inverse-Wishart prior, leverages labeled data to calibrate a shared prior and warm-start posteriors for known categories, and dynamically compares posterior predictive evidence for assigning a sample to an existing class versus creating a new one, adaptively adjusting decision confidence. Evaluated on standard open-category discovery benchmarks, the approach significantly outperforms strong baselines, demonstrating superior performance in discovering novel categories while maintaining competitive accuracy on known classes.

Abstract

On-the-fly category discovery requires deciding for each incoming test sample whether to assign it to an existing category or spawn a new one. Existing methods typically implement this decision through matching-based heuristics, such as radius- or hash-based rules. While effective in practice, these methods usually treat category birth implicitly as a fallback when no existing category matches confidently, rather than as an explicit alternative supported by its own statistical evidence. To address this, we propose DP-BOA, a posterior-predictive decision framework based on an online Dirichlet-process Gaussian mixture model with a Normal-Inverse-Wishart prior. During training, we use labeled data to calibrate a shared NIW prior over category Gaussians and warm-start the known-category posteriors. At test time, for each incoming sample, DP-BOA compares the posterior predictive evidence for assignment to existing categories against the evidence for spawning a new category induced by the DP prior, and then updates category statistics online after the decision. The method captures anisotropic category geometry and naturally adapts decision confidence as evidence accumulates. Across standard OCD benchmarks, DP-BOA consistently outperforms strong baselines and delivers particularly strong novel-class discovery performance while maintaining competitive known-class accuracy.

Problem

• on-the-fly category discovery
• category assignment
• novel class discovery
• Dirichlet process
• posterior predictive

Innovation

• Dirichlet Process
• Online Category Discovery
• Posterior Predictive Decision
• Gaussian Mixture Model
• Normal-Inverse-Wishart Prior