# CAO: Curvature-Adaptive Optimization via Periodic Low-Rank Hessian Sketching

arXiv: https://arxiv.org/abs/2511.12548 · HTML: https://mlsanity.com/paper/2511.12548/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 25 references that are in the map.
- First posted on arXiv 16 Nov 2025.

## Builds on (most cited first, 10 of 10)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Optimizing Neural Networks with Kronecker-factored Approximate Curvature](https://mlsanity.com/paper/1503.05671/) (2015)
- [Linear Convergence of Gradient and Proximal-Gradient Methods Under the Polyak-Lojasiewicz Condition](https://mlsanity.com/paper/1608.04636/) (2016)
- [Shampoo: Preconditioned Stochastic Tensor Optimization](https://mlsanity.com/paper/1802.09568/) (2018)
- [An Investigation into Neural Net Optimization via Hessian Eigenvalue Density](https://mlsanity.com/paper/1901.10159/) (2019)
- [Empirical Analysis of the Hessian of Over-Parametrized Neural Networks](https://mlsanity.com/paper/1706.04454/) (2017)
- [How to Escape Saddle Points Efficiently](https://mlsanity.com/paper/1703.00887/) (2017)
- [Escaping From Saddle Points --- Online Stochastic Gradient for Tensor Decomposition](https://mlsanity.com/paper/1503.02101/) (2015)
- [ADAHESSIAN: An Adaptive Second Order Optimizer for Machine Learning](https://mlsanity.com/paper/2006.00719/) (2020)

## Abstract (arXiv)

First-order optimizers are reliable but slow in sharp, anisotropic regions. We study a curvature-adaptive method that periodically sketches a low-rank Hessian subspace via Hessian--vector products and preconditions gradients only in that subspace, leaving the orthogonal complement first-order. For L-smooth non-convex objectives, we recover the standard O(1/T) stationarity guarantee with a widened stable stepsize range; under a Polyak--Lojasiewicz (PL) condition with bounded residual curvature outside the sketch, the loss contracts at refresh steps. On CIFAR-10/100 with ResNet-18/34, the method enters the low-loss region substantially earlier: measured by epochs to a pre-declared train-loss threshold (0.75), it reaches the threshold 2.95x faster than Adam on CIFAR-100/ResNet-18, while matching final test accuracy. The approach is one-knob: performance is insensitive to the sketch rank k across {1,3,5}, and k=0 yields a principled curvature-free ablation. We release anonymized logs and scripts that regenerate all figures and tables.