The paper introduces a model‑free, noise‑aware Bayesian clustering framework that builds a martingale posterior over clusterings without specifying a likelihood or prior, avoiding costly MCMC and variational methods. It handles irregular cluster shapes and allows parallel posterior sampling. To interpret the high‑dimensional posterior, a Wasserstein‑type summarization compresses many draws into a small set of weighted representatives using a novel noise‑aware loss, while marginalising uncertainty in the density level and providing theoretical guarantees.
By Qing Li, Sara Wade
arXiv:2604. 23083v2 Announce Type: replace-cross Abstract: Generative approaches to clustering provide information on geometric properties of clusters, whereas discriminative approaches provide boundaries between clusters.
By Mackenzie R. Neal, Paul D. McNicholas, Arthur White
The paper formalizes a geometric tradeoff between ambient separation and sampling gaps to determine when distinct manifold components can be reliably separated in clustering. It introduces a threshold phenomenon for mutual‑k‑nearest‑neighbor graphs, defining an uncertainty zone where the number of clusters cannot be identified. The authors propose Manifold‑Based Clustering (MBC), which outputs a bracket interval quantifying this uncertainty rather than forcing a single cluster count.
By Savik Kinger, Luciano Dyballa, Steven W. Zucker
arXiv:2606. 19643v1 Announce Type: cross Abstract: Motivated by the privacy, sensitivity and sharing limitations of health data, we present a comprehensive pipeline for inference of Bayesian mixture models within a federated learning setting, i.
By Julie Fendler, Francesca L. Crowe, Tom Marshall, Sylvia Richardson, Paul D. W. Kirk
arXiv:2606. 00327v1 Announce Type: cross Abstract: Clustering is widely used across the sciences as the foundation for downstream data-driven scientific discoveries.
By Kai R. Wycik, Tiffany M. Tang, Tarek M. Zikry, Genevera I. Allen
arXiv:2602. 24201v2 Announce Type: replace Abstract: Estimating density ratios between pairs of intractable data distributions is a core problem in probabilistic modeling, enabling principled comparisons of sample likelihoods under different data-generating processes across conditions.
By Egor Antipov, Alessandro Palma, Lorenzo Consoli, Stephan G\"unnemann, Andrea Dittadi, Fabian J. Theis
Product Monte Carlo (PMC) is an algorithm that estimates expectations under a product of densities using only Monte Carlo samples from each individual density. By reformulating the intractable product expectation as a regression problem involving computable conditional expectations, PMC provides both point estimates and coherent Bayesian uncertainty quantification. The method achieves higher accuracy, better-calibrated uncertainty, and improved scalability compared to existing techniques, and it does not require density evaluations, enabling its use in simulation‑based inference such as protein image analysis.
By Pietro Carlotti, Sifan Liu, Yuling Yao
arXiv:2608. 12027v1 Announce Type: cross Abstract: Pairwise constrained clustering typically relies on hard must-link/cannot-link labels, whereas realistic pairwise supervision may be real-valued and entangle intrinsic ambiguity, expert judgment, and stochastic corruption.
By Shaojie Zhang, Ke Chen
The paper introduces a general framework for multi-domain clustering using measure quantization, where a shared set of cluster prototypes is learned by minimizing a probability metric (e.g., Sinkhorn divergence or Maximum Mean Discrepancy) between each domain’s probability measure and the prototype measure. Data points are assigned to clusters either by nearest centroid or via optimal transport, a collaborative strategy that couples all samples within a domain. A mini‑batch optimization strategy makes both fitting and assignment scalable, reducing memory and computational cost while preserving clustering performance, and experimental results on five multi‑domain benchmarks (image, audio, and sensor data) show that the Sinkhorn‑based method consistently outperforms classical and multi‑domain clustering baselines, even when scaling to hundreds of thousands of samples.
By Rafael Pereira Eufrazio, Eduardo Fernandes Montesuma, Charles Casimiro Cavalcante
arXiv:2604. 18801v2 Announce Type: replace Abstract: Scientific particle simulations in cosmology, molecular dynamics, and fluid dynamics produce large-scale datasets whose storage, movement, and analysis increasingly rely on lossy compression.
By Congrong Ren, Sheng Di, Katrin Heitmann, Franck Cappello, Hanqi Guo
The paper introduces improvements to the VampPrior, a prior used in deep latent variable models. By decoupling prior and posterior covariances and replacing fixed uniform mixture weights with a Dirichlet process, the authors achieve higher model evidence and better non‑parametric clustering. Applying the enhanced prior to the single‑cell Variational Inference framework improves integration and clustering across multiple atlases while maintaining a continuous latent space.
By Andrew A. Stirn, Abhejit Rajagopal, David A. Knowles, Grace Huynh
arXiv:2606. 00302v1 Announce Type: cross Abstract: Despite being ubiquitous in science, clustering remains a technique whose results are not quantitatively scrutinized via a framework.
By Siamak K. Sorooshyari, Manuel A. Rivas, Robert Tibshirani