Sharp concentration of simple random tensors
Omar Al-Ghattas, Jiaheng Chen*, Daniel Sanz-Alonso · 2025
Read Sharp concentration of simple random tensors by Omar Al-Ghattas, Jiaheng Chen*, Daniel Sanz-Alonso (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-2502-16916.txt.
Description
Omar Al-Ghattas, Jiaheng Chen, and Daniel Sanz-Alonso prove sharp limits on how far sums of random tensors can stray from their average. Tensors are multi-way arrays of numbers, and random versions appear across data science. Their 2025 paper seeks bounds that stay tight regardless of the array's size, a property called dimension-free.
The authors derive concentration bounds and expectation bounds for the deviation of a sum of simple random tensors from its mean. A core step uses generic chaining, a tool for capping the peaks of random processes, to bound families of L_p empirical processes. In doing so they extend classic results, long known for quadratic and product processes, to higher-order settings. The work gives theorists sharper tools for any problem where many random multi-way pieces add up, from statistics to machine learning theory.
Catalog record
| CREATOR | Omar Al-Ghattas, Jiaheng Chen*, Daniel Sanz-Alonso |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 97 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-16916