Any-Class Presence Likelihood for Robust Multi-Label Classification with Abundant Negative Data

2025-06-06

AI Summary

In multi-label classification, the prevalence of negative samples—instances with no positive labels—degrades model discriminability for positive labels. To address this, we propose the Arbitrary-Class-Existence (ACE) Likelihood, which explicitly models the implicit “at-least-one-positive” constraint via normalized weighted geometric averaging of per-class prediction probabilities. We further introduce the first ACE loss, incorporating an adaptive regularization term that balances contributions from absent classes—without additional parameters or computational overhead. Our method is architecture- and loss-agnostic, seamlessly integrating with standard backbones (e.g., ResNet, DenseNet) and base losses (e.g., BCE, ASL). Extensive experiments on SewerML, modified COCO, and ChestX-ray14 demonstrate consistent improvements: up to +6.01 in F1, +8.06 in F2, and +3.11 in mAP. ACE significantly enhances robustness in negative-dominated regimes. Code is publicly available.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Multi-modal Vision
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• 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

Multi-label Classification (MLC) assigns an instance to one or more non-exclusive classes. A challenge arises when the dataset contains a large proportion of instances with no assigned class, referred to as negative data, which can overwhelm the learning process and hinder the accurate identification and classification of positive instances. Nevertheless, it is common in MLC applications such as industrial defect detection, agricultural disease identification, and healthcare diagnosis to encounter large amounts of negative data. Assigning a separate negative class to these instances further complicates the learning objective and introduces unnecessary redundancies. To address this challenge, we redesign standard MLC loss functions by deriving a likelihood of any class being present, formulated by a normalized weighted geometric mean of the predicted class probabilities. We introduce a regularization parameter that controls the relative contribution of the absent class probabilities to the any-class presence likelihood in positive instances. The any-class presence likelihood complements the multi-label learning by encouraging the network to become more aware of implicit positive instances and improve the label classification within those positive instances. Experiments on large-scale datasets with negative data: SewerML, modified COCO, and ChestX-ray14, across various networks and base loss functions show that our loss functions consistently improve MLC performance of their standard loss counterparts, achieving gains of up to 6.01 percentage points in F1, 8.06 in F2, and 3.11 in mean average precision, all without additional parameters or computational complexity. Code available at: https://github.com/ML-for-Sensor-Data-Western/gmean-mlc

Problem

• Handles large negative data in multi-label classification
• Improves identification of positive instances without negative class
• Redesigns loss functions to boost classification performance

Innovation

• Redesigned MLC loss functions with any-class likelihood
• Normalized weighted geometric mean for class probabilities
• Regularization parameter controls absent class contributions