AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Improving Conditional Coverage via Orthogonal Quantile Regression

Shai Feldman, Stephen Bates, Yaniv Romano

NeurIPS 2021 · Learning theory and algorithms › Prediction and decision losses · approach: Probabilistic calibration · 5 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 decorrelation penalty (orthogonal QR)vanilla QR · Synthetic two-group, low noiseminority coverage (%): 66.34 -> 82.55Improved
add decorrelation penalty (orthogonal QR)vanilla QR · Synthetic two-group, high noiseminority coverage (%): 68.74 -> 83.74Improved
add decorrelation penalty (orthogonal QR)vanilla QR · Synthetic two-group, low noiseDelta WSC improvement (%): -13.88Hurt
add decorrelation penalty (orthogonal QR)vanilla QR · Synthetic two-group, high noisePearson correlation improvement (%): +72.21Improved
add decorrelation penalty (orthogonal QR)vanilla QR · facebook_1coverage (%): 88.10 -> 90.48Improved
add decorrelation penalty (orthogonal QR)vanilla QR · blog_datacoverage (%): 82.89 -> 88.92Improved
add decorrelation penalty (orthogonal QR)vanilla QR · navalcoverage (%): 89.89 -> 89.50No clear effect
add decorrelation penalty (orthogonal QR)vanilla QR · nine real datasetsPearson correlation, HSIC: +27.54 to +91.27Improved
add decorrelation penalty (orthogonal QR)vanilla QR · bio, navalDelta WSC improvement (%): -44.73, -57.52Hurt
add decorrelation penalty (orthogonal QR)vanilla QR · kin8nm, navalDelta Node improvement (%): -16.44, -14.34Hurt
add decorrelation penalty (orthogonal QR)vanilla QR · navalinterval length: 0.56 -> 1.49Hurt
conformalize intervals (COQR vs CQR) (ablation)orthogonal QR · real data with conformalizationmarginal coverage (%): 90.11 vs 90.08Improved

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