Publikation
HAD-QC: A Hybrid AI Approach for Automated Quality Control of Argo Float Data
Shivshankar Aiwale; Frederic Theodor Stahl; Lily Sun
In: Artificial Intelligence XLII (SGAI-AI 2025). SGAI International Conference on Innovative Techniques and Applications of Artificial Intelligence (AI-2025), December 16-18, Cambridge, United Kingdom, Lecture Notes in Computer Science (LNAI), Vol. 16302, Springer Nature Switzerland, Cham, 12/2025.
Zusammenfassung
This chapter introduces HAD-QC, a hybrid AI approach designed to enhance the quality control of Argo float data. The text delves into the challenges of real-time and delayed-mode quality control, emphasizing the need for scalable and accurate solutions as the volume and complexity of oceanographic data grow. The core of the chapter focuses on the methodology behind HAD-QC, which combines autoencoder-based anomaly detection, ensemble classification, and traditional rule-based quality control. The evaluation section demonstrates the superior performance of HAD-QC compared to existing methods, showcasing its high precision, recall, and overall accuracy. The chapter also discusses the generalisation of HAD-QC across different float types and ocean basins, highlighting its robustness and potential for widespread operational deployment. Future work and conclusions underscore the importance of integrating HAD-QC into existing data management pipelines and its potential to revolutionize quality control in oceanographic research.