# A New Multipoint Symmetric Secant Method with a Dense Initial Matrix

arXiv: https://arxiv.org/abs/2107.06321 · HTML: https://mlsanity.com/paper/2107.06321/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Jul 2021.

## Built on by (most cited first, 4 of 4)

- [A Machine Learning Analysis of COVID-19 Mental Health Data](https://mlsanity.com/paper/2112.00227/) (2021)
- [A machine learning analysis of the relationship between some underlying medical conditions and COVID-19 susceptibility](https://mlsanity.com/paper/2112.12901/) (2021)
- [Artificial Intelligence-Based Analytics for Impacts of COVID-19 and Online Learning on College Students' Mental Health](https://mlsanity.com/paper/2202.07441/) (2022)
- [A Machine Learning Analysis of Impact of the Covid-19 Pandemic on Alcohol Consumption Habit Changes Among Healthcare Workers in the U.S](https://mlsanity.com/paper/2112.06261/) (2021)

## Abstract (arXiv)

In large-scale optimization, when either forming or storing Hessian matrices are prohibitively expensive, quasi-Newton methods are often used in lieu of Newton's method because they only require first-order information to approximate the true Hessian. Multipoint symmetric secant (MSS) methods can be thought of as generalizations of quasi-Newton methods in that they attempt to impose additional requirements on their approximation of the Hessian. Given an initial Hessian approximation, MSS methods generate a sequence of possibly-indefinite matrices using rank-2 updates to solve nonconvex unconstrained optimization problems. For practical reasons, up to now, the initialization has been a constant multiple of the identity matrix. In this paper, we propose a new limited-memory MSS method for large-scale nonconvex optimization that allows for dense initializations. Numerical results on the CUTEst test problems suggest that the MSS method using a dense initialization outperforms the standard initialization. Numerical results also suggest that this approach is competitive with both a basic L-SR1 trust-region method and an L-PSB method.