A Sobering Look at Tabular Data Generation via Probabilistic Circuits
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911541410922496 |
|---|---|
| author | Scassola, Davide Ponsford, Dylan Javaloy, Adrián Saccani, Sebastiano Bortolussi, Luca Gouk, Henry Vergari, Antonio |
| author_facet | Scassola, Davide Ponsford, Dylan Javaloy, Adrián Saccani, Sebastiano Bortolussi, Luca Gouk, Henry Vergari, Antonio |
| contents | Tabular data is more challenging to generate than text and images, due to its heterogeneous features and much lower sample sizes. On this task, diffusion-based models are the current state-of-the-art (SotA) model class, achieving almost perfect performance on commonly used benchmarks. In this paper, we question the perception of progress for tabular data generation. First, we highlight the limitations of current protocols to evaluate the fidelity of generated data, and advocate for alternative ones. Next, we revisit a simple baseline -- hierarchical mixture models in the form of deep probabilistic circuits (PCs) -- which delivers competitive or superior performance to SotA models for a fraction of the cost. PCs are the generative counterpart of decision forests, and as such can natively handle heterogeneous data as well as deliver tractable probabilistic generation and inference. Finally, in a rigorous empirical analysis we show that the apparent saturation of progress for SotA models is largely due to the use of inadequate metrics. As such, we highlight that there is still much to be done to generate realistic tabular data. Code available at https://github.com/april-tools/tabpc. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_23016 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Sobering Look at Tabular Data Generation via Probabilistic Circuits Scassola, Davide Ponsford, Dylan Javaloy, Adrián Saccani, Sebastiano Bortolussi, Luca Gouk, Henry Vergari, Antonio Machine Learning Artificial Intelligence Tabular data is more challenging to generate than text and images, due to its heterogeneous features and much lower sample sizes. On this task, diffusion-based models are the current state-of-the-art (SotA) model class, achieving almost perfect performance on commonly used benchmarks. In this paper, we question the perception of progress for tabular data generation. First, we highlight the limitations of current protocols to evaluate the fidelity of generated data, and advocate for alternative ones. Next, we revisit a simple baseline -- hierarchical mixture models in the form of deep probabilistic circuits (PCs) -- which delivers competitive or superior performance to SotA models for a fraction of the cost. PCs are the generative counterpart of decision forests, and as such can natively handle heterogeneous data as well as deliver tractable probabilistic generation and inference. Finally, in a rigorous empirical analysis we show that the apparent saturation of progress for SotA models is largely due to the use of inadequate metrics. As such, we highlight that there is still much to be done to generate realistic tabular data. Code available at https://github.com/april-tools/tabpc. |
| title | A Sobering Look at Tabular Data Generation via Probabilistic Circuits |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2603.23016 |