SPACR: Single-Pass Adaptive Training of Uncertainty-Aware Conformal Regressors

Name: SPACR: Single-Pass Adaptive Training of Uncertainty-Aware Conformal Regressors
Uploaded: 2026-09-30T19:32:36.475539+00:00
Duration: 292 s
Description: The episode discusses SPACR: Single-Pass Adaptive Training of Uncertainty-Aware Conformal Regressors by Messoudi soundouss and Rousseau sylvain. The hosts explain that SPACR is a novel method for training uncertainty-aware regression models directly within a differentiable loss function, optimizing efficiency and validity in a single training pass without needing post hoc retraining.

The episode discusses SPACR: Single-Pass Adaptive Training of Uncertainty-Aware Conformal Regressors by Messoudi soundouss and Rousseau sylvain. The hosts explain that SPACR is a novel method for training uncertainty-aware regression models directly within a differentiable loss function, optimizing efficiency and validity in a single training pass without needing post hoc retraining.

Open the full episode