AI Paper Atlas

Research map / Learning theory and algorithms / Optimal transport

Tree-sliced Sobolev IPM

Viet-Hoang Tran, Thanh Tran, Thanh Chu, Duy-Tung Pham, Trung-Khang Tran, Tam Le, Tan Minh Nguyen

ICLR 2026 · Learning theory and algorithms › Optimal transport · approach: Sliced Wasserstein variants · 0 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 TS-Sobolev (p=1.2) instead of Db-TSW in gradient flowTSW · 8 GaussiansWasserstein distance: 2.50e-6 -> 8.88e-7Improved
add TS-Sobolev (p=1.2) instead of Db-TSWTSW · Gaussian 30dWasserstein distance: 1.78 -> 1.40Improved
raise order from p=1.2 to p=2 (ablation)TS-Sobolev · Gaussian 30d gradient flowWasserstein distance: 1.40 -> 3.68Hurt
raise order from p=1.2 to p=2 (ablation)TS-Sobolev · 8 Gaussians gradient flowWasserstein distance: 8.88e-7 -> 3.50e-6Hurt
replace Db-TSW-DD⊥ discriminator loss with TS-Sobolev2-DDDDGAN · CIFAR-10FID: 2.53 -> 2.277Improved
raise order from p=1.5 to p=2 (ablation)TS-Sobolev-DD · CIFAR-10FID: 2.302 -> 2.277Improved
replace Gaussian-kernel uniformity with STS-Sobolev2 instead of STSWSSL · CIFAR-10 projected featuresaccuracy: 76.78 -> 77.65Improved
raise order from p=1.5 to p=2 (ablation)STS-Sobolev · CIFAR-10 SSL encoded featuresaccuracy: 79.88 -> 80.6Improved
use STS-Sobolev2 instead of STSWspherical gradient flow · 12 vMFs epoch 250log Wasserstein distance: -3.549 -> -3.616Improved
use TS-Sobolev2 instead of TSW-SLtopic model · BBCtopic coherence: 0.796 -> 0.805Improved
use STS-Sobolev2 instead of STSWtopic model · M10topic coherence: 0.373 -> 0.423Improved
increase number of trees L (ablation)TS-Sobolev · Monte Carlo approximationapproximation errorImproved

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