AI Paper Atlas

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

Conformal Prediction Sets for Ordinal Classification

PRASENJIT DEY, Srujana Merugu, Sivaramakrishnan R Kaveri

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 2 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
add unimodal construction COPOC instead of vanilla CEV-CE · HCI|PS| size: 3.28 -> 2.66Improved
add unimodal construction COPOC instead of vanilla CEV-CE · HCIAcc@3: 88.9 -> 91.7Improved
enforce unimodality via loss or soft labels (SORD, AVDL, Uni-loss, POE) instead of by constructionbaselines · image datasetscontiguity violations CV%: 24.4 -> 2.1 (V-CE vs AVDL), COPOC 0Hurt
use COPOC instead of SORD on equi-spaced Double-Exponential dataCOPOC · Synthetic D1KL divergence: 0.03 -> 0.04No clear effect
use COPOC instead of V-CE on non-equi-spaced and complex dataCOPOC · Synthetic D4KL divergence: 0.14 -> 0.08Improved
use COPOC instead of V-CE with smaller training setsCOPOC · Synthetic D4 varying NKL divergenceImproved
use LAC instead of APS conformal inferenceCOPOC · HCI|PS| size: 2.66 -> 2.61Improved
use LAC instead of APS conformal inferenceCOPOC · Retina-MNISTSSCV: 0.09 -> 0.24Hurt
use COPOC instead of Ordinal-APSOrdinal-APS · Adience|PS| size: 2.67 -> 2.26Improved
use Ordinal-APS instead of APS on V-CEV-CE · Adience|PS| size: 4.82 -> 2.67Improved
use COPOC instead of V-CE for likelihood fitCOPOC · AestheticNLL: 1.49 -> 0.71Improved
use COPOC instead of V-CE for likelihood fitCOPOC · Retina MNISTNLL: 1.24 -> 1.23No clear effect

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