Scoring rules and their approximations on manifolds

Add to your list(s) Download to your calendar using vCal

• Karthik Bharath (University of Nottingham)
• Friday 06 June 2025, 12:15-12:45
• Seminar Room 1, Newton Institute.

If you have a question about this talk, please contact nobody.

RCLW02 - Calibrating prediction uncertainty : statistics and machine learning perspectives

On metric spaces of strong negative type an energy or kernel-based strictly proper scoring rule for probabilistic forecasts may be defined. However, the relationship between the strong negative type property and the curvature of a metric space that is a manifold is not well understood. I will comment on this issue while drawing parallels to conditions on the curvature that determine efficient sampling on manifolds using intrinsic stochastic differential equations (SDEs). I will then discuss error bounds for SDE -based sampling from forecasts distributions on manifolds, and their application to computing the corresponding scoring rules. 

This talk is part of the Isaac Newton Institute Seminar Series series.

This talk is included in these lists:

• All CMS events
• Featured lists
• INI info aggregator
• Isaac Newton Institute Seminar Series
• School of Physical Sciences
• Seminar Room 1, Newton Institute
• bld31

Note that ex-directory lists are not shown.