# GSVD for Geometry-Grounded Dataset Comparison: An Alignment Angle Is All You Need

arXiv: https://arxiv.org/abs/2603.10283 · HTML: https://mlsanity.com/paper/2603.10283/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 3 of its 23 references that are in the map.
- First posted on arXiv 10 Mar 2026.

## Builds on (most cited first, 3 of 3)

- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [Computational Optimal Transport](https://mlsanity.com/paper/1803.00567/) (2018)
- [Insights on representational similarity in neural networks with canonical correlation](https://mlsanity.com/paper/1806.05759/) (2018)

## Abstract (arXiv)

Geometry-grounded learning asks models to respect structure in the problem domain rather than treating observations as arbitrary vectors. Motivated by this view, we revisit a classical but underused primitive for comparing datasets: linear relations between two data matrices, expressed via the co-span constraint Ax = By = z in a shared ambient space. To operationalize this comparison, we use the generalized singular value decomposition (GSVD) as a joint coordinate system for two subspaces. In particular, we exploit the GSVD form A = HCU, B = HSV with CtopC + StopS = I, which separates shared versus dataset-specific directions through the diagonal structure of (C, S). From these factors we derive an interpretable *angle score* θ(z) ∈ [0, π/2] for a sample z, quantifying whether z is explained relatively more by A, more by B, or comparably by both. The primary role of θ(z) is as a *per-sample geometric diagnostic*. We illustrate the behavior of the score on MNIST through angle distributions and representative GSVD directions. A binary classifier derived from θ(z) is presented as an illustrative application of the score as an interpretable diagnostic tool.