Mengnan Du

Mengnan Du is an Assistant Professor in Data Science at New Jersey Institute of Technology (NJIT). Their research focuses on artificial intelligence, machine learning, and natural language processing with a strong emphasis on fairness, explainability, and bias mitigation in AI systems. Key areas include large language models (LLMs), neural network interpretability, and ethical AI.

Dr. Du leads the Collaborative Research: III: Medium: Towards Effective Detection and Mitigation for Shortcut Learning project funded by the National Science Foundation (2023-2027), exploring framework development for addressing shortcut learning in deep neural networks. They are also involved in interdisciplinary work combining AI with solar physics and financial systems.

Research themes include: 1) XAI (explainable AI) techniques for LLMs, 2) bias detection/mitigation in NLP systems, 3) robustness in neural networks, and 4) AI applications in healthcare/finance. Notable contributions include fairness-aware classifier generalization methods and novel approaches for debiasing knowledge graphs.

Recent work has addressed topics like concept depth analysis in LLMs, generative AI ethics in advertising, and neural operator networks for scientific computing. Their publications span top conferences like ACL, COLING, and WWW, with a focus on bridging theory and real-world AI system development.