# How robust accuracy suffers from certified training with convex relaxations

arXiv: https://arxiv.org/abs/2306.06995 · HTML: https://mlsanity.com/paper/2306.06995/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 32 of its 43 references that are in the map.
- First posted on arXiv 12 Jun 2023.

## Builds on (most cited first, 30 of 32)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [Towards Deep Learning Models Resistant to Adversarial Attacks](https://mlsanity.com/paper/1706.06083/) (2017)
- [Intriguing properties of neural networks](https://mlsanity.com/paper/1312.6199/) (2013)
- [Theoretically Principled Trade-off between Robustness and Accuracy](https://mlsanity.com/paper/1901.08573/) (2019)
- [Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks](https://mlsanity.com/paper/2003.01690/) (2020)
- [Certified Adversarial Robustness via Randomized Smoothing](https://mlsanity.com/paper/1902.02918/) (2019)
- [Robustness May Be at Odds with Accuracy](https://mlsanity.com/paper/1805.12152/) (2018)
- [Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks](https://mlsanity.com/paper/1702.01135/) (2017)
- [Evasion Attacks against Machine Learning at Test Time](https://mlsanity.com/paper/1708.06131/) (2017)
- [Provable defenses against adversarial examples via the convex outer adversarial polytope](https://mlsanity.com/paper/1711.00851/) (2017)
- [Certified Robustness to Adversarial Examples with Differential Privacy](https://mlsanity.com/paper/1802.03471/) (2018)
- [Efficient Neural Network Robustness Certification with General Activation Functions](https://mlsanity.com/paper/1811.00866/) (2018)
- [The Pitfalls of Simplicity Bias in Neural Networks](https://mlsanity.com/paper/2006.07710/) (2020)
- [Certified Defenses against Adversarial Examples](https://mlsanity.com/paper/1801.09344/) (2018)
- [Towards Fast Computation of Certified Robustness for ReLU Networks](https://mlsanity.com/paper/1804.09699/) (2018)
- [Uniform convergence may be unable to explain generalization in deep learning](https://mlsanity.com/paper/1902.04742/) (2019)
- [Towards Stable and Efficient Training of Verifiably Robust Neural Networks](https://mlsanity.com/paper/1906.06316/) (2019)
- [Certified Adversarial Robustness with Additive Noise](https://mlsanity.com/paper/1809.03113/) (2018)
- [Understanding and Mitigating the Tradeoff Between Robustness and Accuracy](https://mlsanity.com/paper/2002.10716/) (2020)
- [Scaling provable adversarial defenses](https://mlsanity.com/paper/1805.12514/) (2018)
- [A Dual Approach to Scalable Verification of Deep Networks](https://mlsanity.com/paper/1803.06567/) (2018)
- [A Convex Relaxation Barrier to Tight Robustness Verification of Neural Networks](https://mlsanity.com/paper/1902.08722/) (2019)
- [Curse of Dimensionality on Randomized Smoothing for Certifiable Robustness](https://mlsanity.com/paper/2002.03239/) (2020)
- [Certified Training: Small Boxes are All You Need](https://mlsanity.com/paper/2210.04871/) (2022)
- [Fast Certified Robust Training with Short Warmup](https://mlsanity.com/paper/2103.17268/) (2021)
- [Adversarial Robustness May Be at Odds With Simplicity](https://mlsanity.com/paper/1901.00532/) (2019)
- [Random Smoothing Might be Unable to Certify ℓ_∞ Robustness for High-Dimensional Images](https://mlsanity.com/paper/2002.03517/) (2020)
- [Why adversarial training can hurt robust accuracy](https://mlsanity.com/paper/2203.02006/) (2022)
- [On the Paradox of Certified Training](https://mlsanity.com/paper/2102.06700/) (2021)

## Built on by (most cited first, 1 of 1)

- [On Using Certified Training towards Empirical Robustness](https://mlsanity.com/paper/2410.01617/) (2024)

## Abstract (arXiv)

Adversarial attacks pose significant threats to deploying state-of-the-art classifiers in safety-critical applications. Two classes of methods have emerged to address this issue: empirical defences and certified defences. Although certified defences come with robustness guarantees, empirical defences such as adversarial training enjoy much higher popularity among practitioners. In this paper, we systematically compare the standard and robust error of these two robust training paradigms across multiple computer vision tasks. We show that in most tasks and for both mathscrl_∞-ball and mathscrl2-ball threat models, certified training with convex relaxations suffers from worse standard and robust error than adversarial training. We further explore how the error gap between certified and adversarial training depends on the threat model and the data distribution. In particular, besides the perturbation budget, we identify as important factors the shape of the perturbation set and the implicit margin of the data distribution. We support our arguments with extensive ablations on both synthetic and image datasets.