mlsanityData to 7 Oct 2026

Geometry-Aware Tabular Diffusion

arXiv 2606.02607 ↗First posted 23 May 2026cs.LG cs.AI cs.CRcited by 0

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Tabular synthesis is critical for privacy-preserving sharing and augmentation, yet diffusion models rely on implicit mechanisms to capture inter-column relationships. We introduce Geometry-Aware Tabular Diffusion (GATD), which augments tabular diffusion denoisers with pairwise angles and lengths computed from column value differences and used as inputs and auxiliary targets. Our MLP instantiation achieves state-of-the-art benchmark performance while using 3.5x fewer parameters on average (up to 25x for classification tasks): on ten datasets, it wins 8/10 Shape, 7/10 Trend, and 9/10 downstream utility (F1/RMSE), reducing Shape and Trend error by 27% and 20%. Default loss weights transfer to GNN and Transformer denoisers, improving Shape on 27/30 and Trend on 25/30 architecture-dataset cells. A matched ablation shows supervision (not extra inputs or capacity) drives the gain. This shows explicit relational supervision is a portable inductive bias for tabular diffusion.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 18 of its 21 references that are in the map. First posted on arXiv 23 May 2026.

Builds on

18 of 21 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024TabDiff: a Mixed-type Diffusion Model for Tabular Data Generation27 Oct
  2. 2023Mixed-Type Tabular Data Synthesis with Score-based Diffusion in Latent Space14 Oct
  3. 2023CoDi: Co-evolving Contrastive Diffusion Models for Mixed-type Tabular Synthesis25 Apr
  4. 2022STaSy: Score-based Tabular data Synthesis8 Oct
  5. 2022TabDDPM: Modelling Tabular Data with Diffusion Models30 Sep

Other references · 13

  1. 2017Attention Is All You Needcited by 59,932
  2. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
  3. 2017Graph Attention Networkscited by 8,854
  4. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  5. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  6. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994
  7. 2016Deep Learning with Differential Privacycited by 2,682
  8. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
  9. 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,131
  10. 2021Revisiting Deep Learning Models for Tabular Datacited by 725
  11. 2019Modeling Tabular data using Conditional GANcited by 695
  12. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 566
1 more
  1. 2016Entity Embeddings of Categorical Variablescited by 85

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025TabRep: Training Tabular Diffusion Models with a Simple and Effective Continuous Representation
  2. 2026CDMD: A Cross-Dataset Mixed-Type Diffusion Model for Tabular Data
  3. 2025Winning the MIDST Challenge: New Membership Inference Attacks on Diffusion Models for Tabular Data Synthesis
  4. 2023AutoDiff: combining Auto-encoder and Diffusion model for tabular data synthesizing
  5. 2025Diffusion and Flow Matching Models for Tabular Data: A Survey
  6. 2026FairDiffuseVQVAE: Sampling-Time Fairness in Tabular Diffusion via Conditional Refinement of Vector-Quantized Latents
  7. 2023Continuous Diffusion for Mixed-Type Tabular Data
  8. 2024ClavaDDPM: Multi-relational Data Synthesis with Cluster-guided Diffusion Models