Zihan Liang 梁梓涵 | zihan.liang@emory.edu | Dept. of Data and Decision Sciences, Emory University

As an Applied Mathematics and Statistics student at Emory University, I am passionate about machine learning, AI, and natural language processing. Through various research projects, I have explored data-driven solutions to real-world challenges in fields like healthcare, autonomous driving, and labor markets. I aim to bridge theoretical research with practical applications, optimizing models to improve data understanding and influence decision-making.

What I'm Doing

My research focuses on Machine Learning (ML), Natural Language Processing (NLP), and Data-Driven Decision Making. I aim to develop robust algorithms and systems that bridge theory and practice, enabling models to handle real-world challenges such as missing data, domain shifts, and multi-modal complexity, with applications spanning healthcare, recommendation systems, and autonomous systems:

• Multimodal Learning under Missingness: designing models that adapt to missing-not-at-random (MNAR) data patterns in clinical, recommendation, and multimodal settings.

• Domain Adaptation and Robust Transfer: developing domain-adaptive frameworks for cross-dataset generalization in computer vision and medical imaging.

• Causal Representation Learning: integrating causal inference with deep learning to build interpretable and generalizable models, especially for healthcare analytics.

• Applied NLP: creating transformer-based architectures for sentiment analysis, clinical text understanding, and health event detection under noisy and imbalanced data.

• AI for Societal Impact: leveraging ML and game-theoretic modeling to inform policy, sustainability, and human-centered AI applications such as autonomous driving and labor markets.