Researchers propose an f-divergence-based certified robustness framework for backdoor poisoning
Start here
A paper published on arXiv on October 26, 2026 proposes an f-divergence-based certified robustness framework for defending against data poisoning and backdoor poisoning attacks. The framework combines data subsampling with noise added to features, labels, and test samples to randomize inputs, and provides a composable certification procedure that can handle any combination of threat model and noise-adding strategy, subsuming several previous certification mechanisms as special cases. The authors state that the experiments provide, for the first time, certification for general ℓp-constrained poisoning adversaries under bagging with Laplace, Gaussian, and Uniform noise.
Generated by AI from reports · Updated 15 days ago
Attention and developments
Attention, hourly
Now 29, 0% in 24 hours·Peak 29 (Nov 6 03:00)
Related research
This event's paper
- A Certified Robustness Framework for Backdoor Poisoning Attacks using $f$-Divergences
arXiv · 2026-10-26 · Deep reading
Progress timeline
- arXivAn f-Divergence-Based Certified Robustness Framework Against Backdoor Poisoning Attacks
Researchers propose a certified robustness framework based on f-divergence for defending against data poisoning and backdoor poisoning attacks. The framework randomizes inputs by combining data subsampling with noise added to features, labels and test samples, and provides a composable certification procedure that can handle any combination of threat model and noise-adding strategy, covering several previous certification mechanisms as special cases. The authors state that the experiments provide, for the first time, certification for poisoning adversaries under general ℓp constraints with bagging combined with Laplace, Gaussian and Uniform noise.