Interpretable intrinsic dimension estimation through componentwise calibration of distance and angle
Abstract
DANCo (Dimensionality from Angle and Norm Concentration) jointly calibrates nearest-neighbor distance and angular statistics and consistently reaches state-of-the-art accuracy on clean intrinsic-dimension (ID) benchmarks. Practical data, however, introduce neighborhood-relative noise and sample-amplitude heterogeneity that can distort these geometric signals. We reformulate DANCo componentwise, retaining separate distance and angular discrepancy curves so that the source of an estimate can be identified and interpreted. For the distance component, we derive a closed-form Kullback-Leibler divergence for the generic-order ratios of the generalized ratios ID estimator (Gride); when both angular parameters are matched (Full), Gride reduces mean percentage error from 27.7% to 17.6% at noise equal to 40% of typical neighbor spacing on 24 manifolds. For the angular component, two sampling regimes motivate aligning mean direction while retaining concentration matching (Profiled). On a Gaussian scale mixture with generating dimension 70 embedded in 100 dimensions, profiling raises the Minimum Neighbor Distance (MiND) estimate from $22.8$ to $66.7$, while removing the known amplitudes restores MiND-Full to $71.9$; the control thus attributes the Full shortfall to amplitude heterogeneity. On CIFAR-10 and ImageNet, amplitude-reducing normalizations move angular location toward the references and narrow the Full-Profiled gap, an observational counterpart to the controlled mixture. Across four pretrained convolutional neural networks, Gride-Profiled, the two-nearest-neighbor estimator (TWO-NN), and the maximum-likelihood estimator (MLE) exhibit similar rise-and-fall profiles, while Full-Profiled differences identify the layers most sensitive to angular calibration.
Related
- Pairwise Distortion Distribution with Graph-Based Importance Sampling for Quantization Error and Intrinsic Dimension Estimation
- Manifold Hypothesis: Conditional Certificates and Consistent Intrinsic Dimension Estimation
- Formalizing the Intrinsic Dimension of Objective Landscapes by a Generalized Notion of Co-dimension
- Intrinsic-dimension empirical Bernstein inequalities for bounded self-adjoint operators
- Minimax Lower Bound for Estimating Diffusion-based Local Intrinsic Dimension
BibTeX
@misc{huang2026interpretable,
title = {{Interpretable intrinsic dimension estimation through componentwise calibration of distance and angle}},
author = {Chih-Hsuan Huang and Chih-Wei Chen and Szu-Chi Chung},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.37114},
eprint = {2609.37114},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.37114},
}
RIS
TY - UNPB
TI - Interpretable intrinsic dimension estimation through componentwise calibration of distance and angle
AU - Chih-Hsuan Huang
AU - Chih-Wei Chen
AU - Szu-Chi Chung
PY - 2026
DA - 2026/09/29
PB - arXiv
DO - 10.48550/arxiv.2609.37114
UR - https://arxiv.org/abs/2609.37114
AB - DANCo (Dimensionality from Angle and Norm Concentration) jointly calibrates nearest-neighbor distance and angular statistics and consistently reaches state-of-the-art accuracy on clean intrinsic-dimension (ID) benchmarks. Practical data, however, introduce neighborhood-relative noise and sample-amplitude heterogeneity that can distort these geometric signals. We reformulate DANCo componentwise, retaining separate distance and angular discrepancy curves so that the source of an estimate can be identified and interpreted. For the distance component, we derive a closed-form Kullback-Leibler divergence for the generic-order ratios of the generalized ratios ID estimator (Gride); when both angular parameters are matched (Full), Gride reduces mean percentage error from 27.7% to 17.6% at noise equal to 40% of typical neighbor spacing on 24 manifolds. For the angular component, two sampling regimes motivate aligning mean direction while retaining concentration matching (Profiled). On a Gaussian scale mixture with generating dimension 70 embedded in 100 dimensions, profiling raises the Minimum Neighbor Distance (MiND) estimate from $22.8$ to $66.7$, while removing the known amplitudes restores MiND-Full to $71.9$; the control thus attributes the Full shortfall to amplitude heterogeneity. On CIFAR-10 and ImageNet, amplitude-reducing normalizations move angular location toward the references and narrow the Full-Profiled gap, an observational counterpart to the controlled mixture. Across four pretrained convolutional neural networks, Gride-Profiled, the two-nearest-neighbor estimator (TWO-NN), and the maximum-likelihood estimator (MLE) exhibit similar rise-and-fall profiles, while Full-Profiled differences identify the layers most sensitive to angular calibration.
ER -