The second edition of the Frontiers in Probabilistic Inference: Learning Meets Sampling (FPI 2025) is a one-day workshop at NeurIPS, focused on advancing scalable, data-efficient sampling methods by connecting classical statistical approaches with modern machine learning techniques. As probabilistic inference plays a growing role in large-scale scientific and real-world systems, this workshop brings together researchers from statistics, machine learning, and applied science domains to explore shared challenges, develop practical tools, and identify common benchmarks. FPI-2025 emphasizes cross-disciplinary collaboration, aiming to foster deeper integration between theory and application in the development of next-generation sampling methods.

We will center workshop discussions around the following topics/questions:

• Sampling methods and their connections to generative models and optimal control.

• Classical sampling approaches and how learning accelerates them.

• Connections between sampling methods and physics.

• Understanding sampling from theoretical perspectives. 

• Applications of sampling to natural sciences, Bayesian inference, LLM fine-tuning, and more.