# Rethinking Benign Overfitting in Two-Layer Neural Networks

arXiv: https://arxiv.org/abs/2502.11893 · HTML: https://mlsanity.com/paper/2502.11893/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 26 references that are in the map.
- First posted on arXiv 17 Feb 2025.

## Builds on (most cited first, 22 of 22)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Neural Tangent Kernel: Convergence and Generalization in Neural Networks](https://mlsanity.com/paper/1806.07572/) (2018)
- [A Convergence Theory for Deep Learning via Over-Parameterization](https://mlsanity.com/paper/1811.03962/) (2018)
- [Gradient Descent Provably Optimizes Over-parameterized Neural Networks](https://mlsanity.com/paper/1810.02054/) (2018)
- [The Implicit Bias of Gradient Descent on Separable Data](https://mlsanity.com/paper/1710.10345/) (2017)
- [What Neural Networks Memorize and Why: Discovering the Long Tail via Influence Estimation](https://mlsanity.com/paper/2008.03703/) (2020)
- [Does Learning Require Memorization? A Short Tale about a Long Tail](https://mlsanity.com/paper/1906.05271/) (2019)
- [An Investigation of Why Overparameterization Exacerbates Spurious Correlations](https://mlsanity.com/paper/2005.04345/) (2020)
- [Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning](https://mlsanity.com/paper/2012.09816/) (2020)
- [Label Noise SGD Provably Prefers Flat Global Minimizers](https://mlsanity.com/paper/2106.06530/) (2021)
- [When is Memorization of Irrelevant Training Data Necessary for High-Accuracy Learning?](https://mlsanity.com/paper/2012.06421/) (2020)
- [Benign Overfitting in Two-layer Convolutional Neural Networks](https://mlsanity.com/paper/2202.06526/) (2022)
- [Shape Matters: Understanding the Implicit Bias of the Noise Covariance](https://mlsanity.com/paper/2006.08680/) (2020)
- [Benign Overfitting without Linearity: Neural Network Classifiers Trained by Gradient Descent for Noisy Linear Data](https://mlsanity.com/paper/2202.05928/) (2022)
- [Implicit Bias of AdamW: ℓ_∞ Norm Constrained Optimization](https://mlsanity.com/paper/2404.04454/) (2024)
- [Towards understanding how momentum improves generalization in deep learning](https://mlsanity.com/paper/2207.05931/) (2022)
- [Benign Overfitting for Two-layer ReLU Convolutional Neural Networks](https://mlsanity.com/paper/2303.04145/) (2023)
- [Memorization Through the Lens of Curvature of Loss Function Around Samples](https://mlsanity.com/paper/2307.05831/) (2023)
- [The Benefits of Mixup for Feature Learning](https://mlsanity.com/paper/2303.08433/) (2023)
- [Memorization in Self-Supervised Learning Improves Downstream Generalization](https://mlsanity.com/paper/2401.12233/) (2024)
- [Memorize to Generalize: on the Necessity of Interpolation in High Dimensional Linear Regression](https://mlsanity.com/paper/2202.09889/) (2022)
- [Measuring Unintended Memorisation of Unique Private Features in Neural Networks](https://mlsanity.com/paper/2202.08099/) (2022)

## Built on by (most cited first, 2 of 2)

- [Understanding the Impact of Differentially Private Training on Memorization of Long-Tailed Data](https://mlsanity.com/paper/2602.03872/) (2026)
- [When Does ℓ2-Boosting Overfit Benignly? High-Dimensional Risk Asymptotics and the ℓ1 Implicit Bias](https://mlsanity.com/paper/2605.06314/) (2026)

## Abstract (arXiv)

Recent theoretical studies (Kou et al., 2023; Cao et al., 2022) have revealed a sharp phase transition from benign to harmful overfitting when the noise-to-feature ratio exceeds a threshold-a situation common in long-tailed data distributions where atypical data is prevalent. However, harmful overfitting rarely happens in overparameterized neural networks. Further experimental results suggested that memorization is necessary for achieving near-optimal generalization error in long-tailed data distributions (Feldman & Zhang, 2020). We argue that this discrepancy between theoretical predictions and empirical observations arises because previous feature-noise data models overlook the heterogeneous nature of noise across different data classes. In this paper, we refine the feature-noise data model by incorporating class-dependent heterogeneous noise and re-examine the overfitting phenomenon in neural networks. Through a comprehensive analysis of the training dynamics, we establish test loss bounds for the refined model. Our findings reveal that neural networks can leverage "data noise" to learn implicit features that improve the classification accuracy for long-tailed data. Our analysis also provides a training-free metric for evaluating data influence on test performance. Experimental validation on both synthetic and real-world datasets supports our theoretical results.