David M. J. Tax

David M. J. Tax is a researcher affiliated with Delft University of Technology, Netherlands. His work spans machine learning, pattern recognition, and computer vision, with a focus on neural networks, anomaly detection, and medical image analysis. He has collaborated extensively with researchers like Marco Loog, Robert P. W. Duin, and Marcel J. T. Reinders on topics including multiple instance learning, dissimilarity-based methods, and physics-informed neural networks.

Key research contributions include advancements in incremental learning for neural networks, personalized anomaly detection in biomedical signals, and frameworks for evaluating time-series anomalies. His recent work (2023–2025) emphasizes physics-informed neural networks, stochastic scheduling algorithms, and proximity-aware evaluation techniques. Tax’s publications reflect a strong emphasis on interdisciplinary applications, integrating computer science with biomedical engineering and operations research.

His methodologies often address challenges in scalability, robustness, and interpretability, with applications ranging from healthcare monitoring to automated project scheduling. Tax maintains an active presence in top-tier conferences like AAAI, ICPR, and NeurIPS, contributing to both theoretical foundations and practical implementations of machine learning systems.