# Why Clipping Matters in AdaGrad? Toward a High-Probability Theory under Generalized Smoothness

arXiv: https://arxiv.org/abs/2609.30276 · HTML: https://mlsanity.com/paper/2609.30276/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 50 of its 61 references that are in the map.
- First posted on arXiv 12 Aug 2026.

## Builds on (most cited first, 30 of 50)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [On the Convergence of Adam and Beyond](https://mlsanity.com/paper/1904.09237/) (2019)
- [Large Batch Optimization for Deep Learning: Training BERT in 76 minutes](https://mlsanity.com/paper/1904.00962/) (2019)
- [Why gradient clipping accelerates training: A theoretical justification for adaptivity](https://mlsanity.com/paper/1905.11881/) (2019)
- [Lower Bounds for Non-Convex Stochastic Optimization](https://mlsanity.com/paper/1912.02365/) (2019)
- [Adaptive Gradient Methods with Dynamic Bound of Learning Rate](https://mlsanity.com/paper/1902.09843/) (2019)
- [AdaGrad stepsizes: Sharp convergence over nonconvex landscapes](https://mlsanity.com/paper/1806.01811/) (2018)
- [A Simple Convergence Proof of Adam and Adagrad](https://mlsanity.com/paper/2003.02395/) (2020)
- [On the Convergence of Stochastic Gradient Descent with Adaptive Stepsizes](https://mlsanity.com/paper/1805.08114/) (2018)
- [Stochastic Polyak Step-size for SGD: An Adaptive Learning Rate for Fast Convergence](https://mlsanity.com/paper/2002.10542/) (2020)
- [A Sufficient Condition for Convergences of Adam and RMSProp](https://mlsanity.com/paper/1811.09358/) (2018)
- [Stochastic Optimization with Heavy-Tailed Noise via Accelerated Gradient Clipping](https://mlsanity.com/paper/2005.10785/) (2020)
- [Convergence of Adam Under Relaxed Assumptions](https://mlsanity.com/paper/2304.13972/) (2023)
- [On the Convergence of Adaptive Gradient Methods for Nonconvex Optimization](https://mlsanity.com/paper/1808.05671/) (2018)
- [Closing the Generalization Gap of Adaptive Gradient Methods in Training Deep Neural Networks](https://mlsanity.com/paper/1806.06763/) (2018)
- [Robustness to Unbounded Smoothness of Generalized SignSGD](https://mlsanity.com/paper/2208.11195/) (2022)
- [High-probability Bounds for Non-Convex Stochastic Optimization with Heavy Tails](https://mlsanity.com/paper/2106.14343/) (2021)
- [A High Probability Analysis of Adaptive SGD with Momentum](https://mlsanity.com/paper/2007.14294/) (2020)
- [The Power of Adaptivity in SGD: Self-Tuning Step Sizes with Unbounded Gradients and Affine Variance](https://mlsanity.com/paper/2202.05791/) (2022)
- [High-Probability Bounds for Stochastic Optimization and Variational Inequalities: the Case of Unbounded Variance](https://mlsanity.com/paper/2302.00999/) (2023)
- [Convergence of AdaGrad for Non-convex Objectives: Simple Proofs and Relaxed Assumptions](https://mlsanity.com/paper/2305.18471/) (2023)
- [High Probability Convergence of Stochastic Gradient Methods](https://mlsanity.com/paper/2302.14843/) (2023)
- [From Gradient Clipping to Normalization for Heavy Tailed SGD](https://mlsanity.com/paper/2410.13849/) (2024)
- [Convex and Non-convex Optimization Under Generalized Smoothness](https://mlsanity.com/paper/2306.01264/) (2023)
- [Beyond Uniform Smoothness: A Stopped Analysis of Adaptive SGD](https://mlsanity.com/paper/2302.06570/) (2023)
- [High Probability Bounds for a Class of Nonconvex Algorithms with AdaGrad Stepsize](https://mlsanity.com/paper/2204.02833/) (2022)
- [Stability and Convergence of Stochastic Gradient Clipping: Beyond Lipschitz Continuity and Smoothness](https://mlsanity.com/paper/2102.06489/) (2021)
- [Generalized-Smooth Nonconvex Optimization is As Efficient As Smooth Nonconvex Optimization](https://mlsanity.com/paper/2303.02854/) (2023)
- [Provable Adaptivity of Adam under Non-uniform Smoothness](https://mlsanity.com/paper/2208.09900/) (2022)
- [SGD with AdaGrad Stepsizes: Full Adaptivity with High Probability to Unknown Parameters, Unbounded Gradients and Affine Variance](https://mlsanity.com/paper/2302.08783/) (2023)

## Abstract (arXiv)

We analyze the original same-step coordinate-wise AdaGrad under generalized smoothness and heavy-tailed noise with bounded variance. In this setting, local curvature may grow sub-quadratically with the gradient norm, and stochastic gradients are assumed to have only bounded conditional second moments. We show that unclipped AdaGrad can become anisotropically miscalibrated: under heavy-tailed noise, the adaptive denominator can learn the geometry of rare noise shocks rather than the local curvature of the objective, leading to a persistent directional distortion that blocks finite-horizon Euclidean progress. We then prove that clipping repairs this failure mode. Our main result is a finite-horizon high-probability guarantee for the original non-lagged AdaGrad update, yielding frac1Tsumt=0T-1|nabla f(xt)|2=Oleft(fracdbig(√log T + log frac1δbig)√Tright), and hence widetildemathcal O(varepsilon-2) complexity. This shows that, for AdaGrad under heavy-tailed noise, clipping is a structural stabilizer of the adaptive geometry rather than merely a robustness heuristic.