Saved in:
Bibliographic Details
Main Authors: Zhao, Tianqi, Wang, Guanyang, Tan, Yan Shuo, Zhang, Qiong
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.21656
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911682752675840
author Zhao, Tianqi
Wang, Guanyang
Tan, Yan Shuo
Zhang, Qiong
author_facet Zhao, Tianqi
Wang, Guanyang
Tan, Yan Shuo
Zhang, Qiong
contents Clustering tabular data is a fundamental yet challenging problem due to heterogeneous feature types, diverse data-generating mechanisms, and the absence of transferable inductive biases across datasets. Prior-fitted networks (PFNs) have recently demonstrated strong generalization in supervised tabular learning by amortizing Bayesian inference under a broad synthetic prior. Extending this paradigm to clustering is nontrivial: clustering is unsupervised, admits a combinatorial and permutation-invariant output space, and requires inferring the number of clusters. We introduce TabClustPFN, a prior-fitted network for tabular data clustering that performs amortized Bayesian inference over both cluster assignments and cluster cardinality. Pretrained on synthetic datasets drawn from a flexible clustering prior, TabClustPFN clusters unseen datasets in a single forward pass, without dataset-specific retraining or hyperparameter tuning. The model naturally handles heterogeneous numerical and categorical features and adapts to a wide range of clustering structures. Experiments on synthetic data and curated real-world tabular benchmarks show that TabClustPFN outperforms classical, deep, and amortized clustering baselines, while exhibiting strong robustness in out-of-the-box exploratory settings. Code is available at https://github.com/Tianqi-Zhao/TabClustPFN.
format Preprint
id arxiv_https___arxiv_org_abs_2601_21656
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TabClustPFN: A Prior-Fitted Network for Tabular Data Clustering
Zhao, Tianqi
Wang, Guanyang
Tan, Yan Shuo
Zhang, Qiong
Machine Learning
Clustering tabular data is a fundamental yet challenging problem due to heterogeneous feature types, diverse data-generating mechanisms, and the absence of transferable inductive biases across datasets. Prior-fitted networks (PFNs) have recently demonstrated strong generalization in supervised tabular learning by amortizing Bayesian inference under a broad synthetic prior. Extending this paradigm to clustering is nontrivial: clustering is unsupervised, admits a combinatorial and permutation-invariant output space, and requires inferring the number of clusters. We introduce TabClustPFN, a prior-fitted network for tabular data clustering that performs amortized Bayesian inference over both cluster assignments and cluster cardinality. Pretrained on synthetic datasets drawn from a flexible clustering prior, TabClustPFN clusters unseen datasets in a single forward pass, without dataset-specific retraining or hyperparameter tuning. The model naturally handles heterogeneous numerical and categorical features and adapts to a wide range of clustering structures. Experiments on synthetic data and curated real-world tabular benchmarks show that TabClustPFN outperforms classical, deep, and amortized clustering baselines, while exhibiting strong robustness in out-of-the-box exploratory settings. Code is available at https://github.com/Tianqi-Zhao/TabClustPFN.
title TabClustPFN: A Prior-Fitted Network for Tabular Data Clustering
topic Machine Learning
url https://arxiv.org/abs/2601.21656