Towards Deep Learning Models Resistant to Adversarial Attacks

Research question

Can robustness be framed as a well-defined optimisation problem?

Key findings

Robust optimisation (min-max) formulation with projected gradient descent as the inner attack - PGD adversarial training gives the first reliably robust MNIST and CIFAR models - Robustness requires larger model capacity

Sources

arxiv.orghttps://arxiv.org/abs/1706.06083

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit research Report a correction or request removal

Related in Research Paper Library

The More It Says, the More You Pay: A Black-Box Audit of Provider-Side Token Inflation in LLM Services

In pay-per-token LLM services, the more a model says, the more users pay. Dishonest providers can covertly manipulate generation to inflate output tokens while largely preserving task utility. We define such…

Adversarial examples and evasion2026

A GAN-Based Framework for Robust DDoS Attack Detection

The availability and consistency of online services remain vulnerable due to Distributed Denial of Service (DDoS) attacks. These attacks are evolving by adopting more complex strategies to evade traditional network…

Adversarial examples and evasion2026

When Agents Look Like Beacons: NIDS Evasion by Model Context Protocol Traffic

The Model Context Protocol (MCP) standardizes communication between autonomous Artificial Intelligence (AI) agents and remote tools over Streamable HTTP. This shift introduces a class of machine-generated,…

Adversarial examples and evasion2026