Isometry pursuit

📰 ArXiv cs.AI

Learn to identify orthonormal column-submatrices using isometry pursuit, a convex algorithm for wide matrices, and apply it to Jacobians for isometric embeddings

advanced Published 28 May 2026

Action Steps

Apply isometry pursuit to a wide matrix to identify orthonormal column-submatrices
Use the novel normalization method to preprocess the matrix
Run multitask basis pursuit on the normalized matrix to obtain the isometric embeddings
Evaluate the performance of isometry pursuit using theoretical and experimental results
Apply isometry pursuit to Jacobians of putative coordinate functions to identify isometric embeddings from within interpretable dictionaries

Who Needs to Know This

Data scientists and ML engineers can benefit from this method for identifying isometric embeddings, which can be useful in various applications such as dimensionality reduction and feature selection

Key Insight

💡 Isometry pursuit can be used to identify isometric embeddings from within interpretable dictionaries, which can be useful in various applications

Key Takeaways

Learn to identify orthonormal column-submatrices using isometry pursuit, a convex algorithm for wide matrices, and apply it to Jacobians for isometric embeddings

Full Article

Title: Isometry pursuit

Abstract:
arXiv:2411.18502v2 Announce Type: replace-cross Abstract: Isometry pursuit is a convex algorithm for identifying orthonormal column-submatrices of wide matrices. It consists of a novel normalization method followed by multitask basis pursuit. Applied to Jacobians of putative coordinate functions, it helps identity isometric embeddings from within interpretable dictionaries. We provide theoretical and experimental results justifying this method. For problems involving coordinate selection and div

Read full paper → ← Back to Reads