ProHOC: Probabilistic Hierarchical Out-of-Distribution Classification via Multi-Depth Networks

2025-03-27

AI Summary

Traditional out-of-distribution (OOD) detection treats samples as either in-distribution (ID) or OOD without considering their semantic hierarchical relationships with known classes. This work proposes the first probabilistic hierarchical OOD classification framework: ID samples are mapped to leaf nodes of a predefined class hierarchy, while OOD samples are assigned to the most semantically appropriate internal node—enabling fine-grained, semantically interpretable OOD identification. Methodologically, we jointly train multi-depth neural networks to construct a hierarchical probabilistic model and perform OOD confidence propagation over the class taxonomy. Evaluated on three benchmark datasets with explicit hierarchical structure, our approach significantly improves both semantic plausibility and localization accuracy of OOD detection. The implementation is publicly available.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Out-of-distribution (OOD) detection in deep learning has traditionally been framed as a binary task, where samples are either classified as belonging to the known classes or marked as OOD, with little attention given to the semantic relationships between OOD samples and the in-distribution (ID) classes. We propose a framework for detecting and classifying OOD samples in a given class hierarchy. Specifically, we aim to predict OOD data to their correct internal nodes of the class hierarchy, whereas the known ID classes should be predicted as their corresponding leaf nodes. Our approach leverages the class hierarchy to create a probabilistic model and we implement this model by using networks trained for ID classification at multiple hierarchy depths. We conduct experiments on three datasets with predefined class hierarchies and show the effectiveness of our method. Our code is available at https://github.com/walline/prohoc.

Problem

• Detects and classifies OOD samples using class hierarchy
• Predicts OOD data to internal nodes in hierarchy
• Leverages multi-depth networks for probabilistic classification

Innovation

• Hierarchical OOD classification via multi-depth networks
• Probabilistic model using class hierarchy relationships
• Leveraging ID classification networks at multiple depths