[Submitted on 29 Jan 2026 (v1), last revised 15 Oct 2026 (this version, v4)]
Abstract:Prior-data Fitted Networks (PFNs) have reframed supervised tabular learning as single-pass in-context inference without per-dataset optimization. Extending this paradigm to clustering is appealing yet fundamentally more challenging, due to absent supervision, unknown cluster cardinality, and label switching inherent to partition outputs. Existing PFN-based clustering methods address these challenges only partially, either requiring known cardinality as input or relying on unstable label-ordering conventions and overly restrictive synthetic priors. We introduce TabClustPFN, a clustering PFN that resolves all challenges jointly through co-designed prior, objective, and architecture. Our hybrid pretraining prior captures heterogeneous real-tabular geometry; our decoupled partition inference network–cardinality inference network architecture jointly infers cluster assignments and cardinality in a single pass; and our SoftARI training objective is permutation-invariant by construction, eliminating the need for any label-ordering convention. On a 44 curated real-world tabular benchmark, TabClustPFN achieves state-of-the-art clustering performance against classical, deep, and amortized baselines, with runtime comparable to efficient baselines. Code is available at this https URL.
Submission history
From: Tianqi Zhao [view email]
[v1] Thu, 29 Jan 2026 12:56:41 UTC (5,185 KB)
[v2] Fri, 30 Jan 2026 07:18:19 UTC (10,372 KB)
[v3] Thu, 14 May 2026 04:07:25 UTC (5,268 KB)
[v4] Thu, 15 Oct 2026 16:09:43 UTC (4,969 KB)
