How Geometry Sheds Light on Statistical Inference 几何如何启发统计推断

This lecture discussed how geometric ideas can provide a useful lens for statistical inference. The talk focused on robust Bayesian inference on manifolds and on the information-geometric meaning of Bayesian updating.

这次讲座讨论几何观点如何为统计推断提供新的理解视角，重点包括流形上的稳健贝叶斯推断， 以及贝叶斯更新在信息几何中的解释。

Topics主题

• Robust Bayesian inference on manifolds.
• Information geometry and Fisher-Rao metric.
• Bayesian updating and exponential families.
• Personal reflections and open questions.