Tianjin Huang

Tianjin Huang is a University Researcher in the Mathematics and Computer Science school at Eindhoven University of Technology, focusing on deep learning, adversarial training, and sparsity. Their work spans traffic prediction, fairness in AI, and robustness evaluation.

• Specializes in Adversarial Machine Learning and Spatiotemporal Graph Modeling
• Develops novel frameworks for Traffic Prediction under real-world data challenges
• Explores Sparse Neural Networks for efficiency and performance

Recent research trends include 4-bit training stability for LLMs, confusional spectral regularization for fairness, and principal eigenvalue methods for certified robustness. Their 2025 publications demonstrate cutting-edge work in adversarial training, graph neural networks, and low-bit optimization.

Scientific recognition includes the Best Paper Award of LoG 2022 for collaborative work on sparse neural network training. They collaborate with institutions across Europe and Asia, focusing on trustworthy AI systems.