Adaptive Gradient Calibration for Single-Positive Multi-Label Learning in Remote Sensing Image Scene Classification

2025-10-09

AI Summary

To address the supervision ambiguity in remote sensing image scene classification caused by single-positive multi-label learning (SPML)—where only one positive label is provided per sample—this paper proposes AdaGC, an adaptive gradient calibration framework. AdaGC integrates gradient calibration, Mixup-based data augmentation, and a dual exponential moving average (EMA) mechanism for pseudo-label generation. Crucially, it introduces a training-dynamics-aware adaptive triggering strategy that initiates calibration only after a warm-up phase to ensure high-quality pseudo-labels. This design effectively mitigates annotation noise interference, suppresses overfitting, and enhances model generalization. Extensive experiments on two benchmark remote sensing datasets demonstrate that AdaGC achieves state-of-the-art performance under two representative annotation noise settings, significantly outperforming existing SPML methods. The results validate AdaGC’s robustness and practical efficacy for noisy weakly supervised remote sensing classification.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Remote Sensing / Geospatial AI
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling

Abstract

Multi-label classification (MLC) offers a more comprehensive semantic understanding of Remote Sensing (RS) imagery compared to traditional single-label classification (SLC). However, obtaining complete annotations for MLC is particularly challenging due to the complexity and high cost of the labeling process. As a practical alternative, single-positive multi-label learning (SPML) has emerged, where each image is annotated with only one relevant label, and the model is expected to recover the full set of labels. While scalable, SPML introduces significant supervision ambiguity, demanding specialized solutions for model training. Although various SPML methods have been proposed in the computer vision domain, research in the RS context remains limited. To bridge this gap, we propose Adaptive Gradient Calibration (AdaGC), a novel and generalizable SPML framework tailored to RS imagery. AdaGC adopts a gradient calibration (GC) mechanism combined with Mixup and a dual exponential moving average (EMA) module for robust pseudo-label generation. To maximize AdaGC's effectiveness, we introduce a simple yet theoretically grounded indicator to adaptively trigger GC after an initial warm-up stage based on training dynamics, thereby guaranteeing the effectiveness of GC in mitigating overfitting to label noise. Extensive experiments on two benchmark RS datasets under two distinct label noise types demonstrate that AdaGC achieves state-of-the-art (SOTA) performance while maintaining strong robustness across diverse settings.

Problem

• Addresses single-positive multi-label learning in remote sensing imagery.
• Mitigates supervision ambiguity in incomplete label annotations.
• Reduces overfitting to label noise through adaptive gradient calibration.

Innovation

• Adaptive Gradient Calibration for SPML learning
• Combines gradient calibration with Mixup and EMA
• Uses training dynamics to adaptively trigger calibration