Yong Lin

Yong Lin is a Postdoctoral Research Associate at the Princeton AI Lab, Princeton University. His work focuses on artificial intelligence, machine learning, and automated theorem proving, with a particular emphasis on improving the robustness, generalization, and alignment of large language models (LLMs).

Research interests include

• Automated theorem proving for formal systems
• Reinforcement learning and reward modeling
• Out-of-distribution generalization in deep learning
• Efficient model merging and parameter control

Recent publications highlight advancements in mathematical reasoning, LLM alignment, and robust AI systems, with a focus on addressing spurious correlations, invariant principles, and compositional reasoning challenges in multimodal and language models.

Yong Lin collaborates with Chi Jin as part of the Princeton AI Lab, contributing to open-source frameworks like Goedel-prover while exploring theoretical foundations for practical AI applications.