A Modern Analysis of Hutchinson's Trace Estimator
This work provides improved theoretical bounds for a fundamental numerical estimation technique, which is significant for researchers and practitioners in numerical linear algebra and machine learning.
This paper improves the accuracy analysis of Hutchinson's trace estimator by leveraging hypercontractive inequalities and concentration properties of sub-gamma distributions. The result is a new state-of-the-art in accuracy analysis with numerically superior bounds.
The paper establishes the new state-of-art in the accuracy analysis of Hutchinson's trace estimator. Leveraging tools that have not been previously used in this context, particularly hypercontractive inequalities and concentration properties of sub-gamma distributions, we offer an elegant and modular analysis, as well as numerically superior bounds. Besides these improvements, this work aims to better popularize the aforementioned techniques within the CS community.