AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

From Classification Accuracy to Proper Scoring Rules: Elicitability of Probabilistic Top List Predictions

Johannes Resin

JMLR 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use true top-1 list instead of mode predictionpadded symmetric scores · 5-class exampleexpected padded Brier score, p(m): 1 -> 0.6875 (70.59% -> 17.28% above optimal)Improved
use true top-2 list instead of top-1 listpadded symmetric scores · 5-class exampleexpected padded Brier score, p(m): 0.6875 -> 0.5867Improved
use true top-2 list instead of top-1 listpadded symmetric scores · 5-class exampleexpected padded logarithmic score, p(m): 1.3863 -> 1.0532Improved
use true top-2 list instead of top-1 listpadded symmetric scores · 5-class exampleexpected padded Brier score, p(l): 0.7969 -> 0.7955No clear effect
use padded logarithmic instead of padded Brier score for top-1 listpadded symmetric scores · 5-class examplerelative deviation from optimal score, p(h): 0.38% -> 24.75%Hurt
use probabilistic top-1 list instead of hard modepadded symmetric scores · 5-class exampleexpected padded logarithmic score: infinity -> 0.0699Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy