AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

TabularBERT: Binning-Based Self-Supervised Learning for Tabular Representation

Beomjin Park, Seunghwan An, Sungchul Hong, Hosik Choi

ICML 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning

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 SWD (Wasserstein) pretraining objective instead of CE (ablation)TabularBERT · ADaccuracy: 0.858 -> 0.862Improved
add SWD (Wasserstein) pretraining objective instead of CE (ablation)TabularBERT · GEaccuracy: 0.715 -> 0.733Improved
add SWD (Wasserstein) pretraining objective instead of CE (ablation)TabularBERT · EYaccuracy: 0.787 -> 0.831Improved
add SWD (Wasserstein) pretraining objective instead of CE (ablation)TabularBERT · CARMSE: 0.440 -> 0.441No clear effect
add SWD (Wasserstein) pretraining objective instead of CE (ablation)TabularBERT · HORMSE: 3.149 -> 3.113Improved
add auxiliary reconstruction loss in pretraining (ablation)TabularBERT · ADaccuracy: 0.860 -> 0.862Improved
add auxiliary reconstruction loss in pretraining (ablation)TabularBERT · GEaccuracy: 0.716 -> 0.733Improved
add auxiliary reconstruction loss in pretraining (ablation)TabularBERT · EYaccuracy: 0.796 -> 0.831Improved
add auxiliary reconstruction loss in pretraining (ablation)TabularBERT · CARMSE: 0.440 -> 0.441No clear effect
add auxiliary reconstruction loss in pretraining (ablation)TabularBERT · HORMSE: 3.151 -> 3.113Improved
increase regularization level together with many bins (ablation)TabularBERT · CH/GE/HOaccuracy/RMSEImproved
use strong fused-type regularization with few bins (ablation)TabularBERT · CH/GE/HOaccuracy/RMSEHurt

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