Selective Inference for Deep Clustering in Latent Spaces
Abstract
Deep clustering is a powerful approach for discovering meaningful structures in high-dimensional data by learning a low-dimensional latent representation prior to clustering. Despite its empirical success, assessing the statistical reliability of the resulting clusters remains challenging. Testing discovered clusters on the same data induces selection bias and invalidates classical p-values. Selective inference (SI) provides a principled framework for correcting this bias, but existing methods focus on clustering performed directly on the observed features. In this work, we develop an SI framework for deep clustering with a fixed pretrained encoder. The key challenge is that cluster assignments are determined through a nonlinear transformation from the original data space to the latent space, resulting in a substantially more complex selection process than in conventional clustering. Our method provides a computationally tractable way to account for this process and enables valid statistical testing of differences between clusters identified in the latent space. Synthetic experiments demonstrate that the proposed method controls the Type I error rate while achieving higher power than valid but conservative baselines, and genomic applications show that it can identify significant cluster differences while appropriately accounting for selection bias. Our framework provides a principled approach to quantifying the statistical reliability of structures discovered by deep clustering.
Related
- Selective Inference for Clustering Under Unknown Covariance
- Exact Selective Inference for Spectral Clustering via Piecewise-Constant Graph Perturbation
- High-Dimensional Simulation-Based Inference in Latent Spaces
- A latent space hidden Markov model for clustering views in longitudinal multilayer networks
- Latent-Class Poisson-Lognormal Mixed Models for Clustering Multivariate Longitudinal Count Data
BibTeX
@misc{mizui2026selective,
title = {{Selective Inference for Deep Clustering in Latent Spaces}},
author = {Eina Mizui and Tomohiro Shiraishi and Shunichi Nishino and Ichiro Takeuchi},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.28756},
eprint = {2609.28756},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.28756},
}
RIS
TY - UNPB
TI - Selective Inference for Deep Clustering in Latent Spaces
AU - Eina Mizui
AU - Tomohiro Shiraishi
AU - Shunichi Nishino
AU - Ichiro Takeuchi
PY - 2026
DA - 2026/09/23
PB - arXiv
DO - 10.48550/arxiv.2609.28756
UR - https://arxiv.org/abs/2609.28756
AB - Deep clustering is a powerful approach for discovering meaningful structures in high-dimensional data by learning a low-dimensional latent representation prior to clustering. Despite its empirical success, assessing the statistical reliability of the resulting clusters remains challenging. Testing discovered clusters on the same data induces selection bias and invalidates classical p-values. Selective inference (SI) provides a principled framework for correcting this bias, but existing methods focus on clustering performed directly on the observed features. In this work, we develop an SI framework for deep clustering with a fixed pretrained encoder. The key challenge is that cluster assignments are determined through a nonlinear transformation from the original data space to the latent space, resulting in a substantially more complex selection process than in conventional clustering. Our method provides a computationally tractable way to account for this process and enables valid statistical testing of differences between clusters identified in the latent space. Synthetic experiments demonstrate that the proposed method controls the Type I error rate while achieving higher power than valid but conservative baselines, and genomic applications show that it can identify significant cluster differences while appropriately accounting for selection bias. Our framework provides a principled approach to quantifying the statistical reliability of structures discovered by deep clustering.
ER -