AI Paper Atlas

Research map / Learning theory and algorithms / Optimal transport

Performance Scaling via Optimal Transport: Enabling Data Selection from Partially Revealed Sources

Feiyang Kang, Hoang Anh Just, Anit Kumar Sahu, Ruoxi Jia

NeurIPS 2023 · Learning theory and algorithms › Optimal transport · approach: Neural optimal transport · 1 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 projektor/PQ instead of Linearperformance predictor · CIFAR-10 extrapolationtest MAE: 6.04 -> 1.26Improved
use projektor/PQ instead of Quadraticperformance predictor · CIFAR-10 extrapolationtest MAE: 5.22 -> 1.26Improved
use projektor/PQ instead of Rationalperformance predictor · ImageNet100 extrapolationtest MAE: 2.27 -> 0.38Improved
use projektor/PQ instead of Quadraticperformance predictor · IMDB extrapolationtest MAE: 2.74 -> 1.16Improved
use projektor/PQ instead of Linearperformance predictor · MNIST extrapolationtest MAE: 28.03 -> 7.27Improved
use projektor/CS instead of projektor/PQ (ablation)performance predictor · MNIST extrapolationtest MAE: 7.27 -> 4.26Improved
use PQ (b,c functions of mixing ratio) instead of CS (ablation)projektor · CIFAR-10, ImageNet100, IMDB extrapolationtest MAE: 1.95 -> 1.26Improved
PQ gives lower training MAE than CS (ablation)projektor · IMDB training fittrain MAE: 0.90 -> 0.25Improved
add OT distance as feature to quadratic-style predictors (projektor/PQ vs Pseudo-Quadratic) (ablation)performance predictor · CIFAR-10 extrapolationtest MAE: 2.90 -> 1.26Improved
fit with more compositions: CS needs about 15 trainings, PQ about 25 to convergeprojektor · CIFAR-10 extrapolationMAE vs number of trainings: 15 vs 25 trainingsImproved
use projektor for mislabeled sources performance projection to 2K-10K from 1Kprojektor/PQ · CIFAR-10 mislabeled sourcesMAE: below 2%Improved
use projektor for unlabeled sources projection to larger scalesprojektor · CIFAR-10 unlabeled sourcesprojection error: below 10%Improved

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