Exploring the Limitations of Structured Orthogonal Dictionary Learning
Anirudh Dash Aditya Siripuram Department of Electrical Engineering Department of Electrical Engineering Indian Institute of Technology, Hyderabad Indian Institute of Technology, Hyderabad [email protected] [email protected] · 2025
Read Exploring the Limitations of Structured Orthogonal Dictionary Learning by Anirudh Dash Aditya Siripuram Department of Electrical Engineering Department of Electrical Engineering Indian Institute of Technology, Hyderabad Indian Institute of Technology, Hyderabad [email protected] [email protected] (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-15094.txt.
Description
Anirudh Dash and Aditya Siripuram probe the limits of learning structured dictionaries from data. Their 2025 paper concerns dictionaries built as products of Householder reflections, a compact way to represent orthogonal transforms used in signal processing. Two questions organize the work: how well can any orthogonal matrix be approximated by a fixed number of such reflections, and how few data samples does learning such a dictionary require?
For the first question, their algorithm outputs an exact decomposition when one exists and bounds the error when it does not. For the second, under a binary coefficient assumption, they prove a striking economy: two data columns suffice to solve the learning problem exactly. The results mark both the power and the boundaries of the structured approach.
Catalog record
| CREATOR | Anirudh Dash Aditya Siripuram Department of Electrical Engineering Department of Electrical Engineering Indian Institute of Technology, Hyderabad Indian Institute of Technology, Hyderabad [email protected] [email protected] |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-15094