AI Paper Atlas

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

JAPAN: Joint Adaptive Prediction Areas with Normalising Flow

Eshant English, Christoph Lippert

ICLR 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Conformal prediction sets

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 flow log-density instead of distance-based conformity scoresconformal prediction · Toy densities (Spiral)prediction area: 25.21 -> 11.65 (vs CONTRA)Improved
use density score vs residual-based CQRconformal prediction · RF4Dprediction area: 1180.65 -> 24.11Improved
use density score vs CONTRAconformal prediction · COVID-19prediction area: 563.64 -> 400.94Improved
use density score vs CONTRAconformal prediction · Droneprediction area: 1.51 -> 1.47No clear effect
use density score vs CONTRAconformal prediction · Energyprediction area: 18.81 -> 16.32No clear effect
train flow conditional on base-model prediction instead of unconditional marginal (ablation)JAPAN · RF4D / Energyprediction area: 170.56 -> 50.10 (Energy)Improved
use posterior p(y-hat|y) score instead of likelihood score (ablation)JAPAN conditional · RF4Dprediction area: 0.69 (x10^3) -> 67.87Hurt
use posterior score instead of likelihood score (ablation)JAPAN conditional · Energyprediction area: 50.10 -> 60.85Hurt
use posterior score instead of likelihood score (ablation)JAPAN · Crescent density (toy)prediction area: 5.57 -> 4.80Improved
threshold latent density only (drop volume correction term) (ablation)JAPAN · Crescent density (toy)coverage/areaHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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