Skip to content

Back to the catalog

For Thirty Pieces of Silver: Overcoming Non-IID via ZKP Sample-Level Auditing and Forcing Attackers into Traitors in Federated Learning
eprint.iacr.orgpaper

For Thirty Pieces of Silver: Overcoming Non-IID via ZKP Sample-Level Auditing and Forcing Attackers into Traitors in Federated Learning

Preprint (IACR ePrint 2026/2611) proposing a federated learning defense that replaces gradient detection with zero-knowledge-proof sample-level audits of client datasets. A staking mechanism forces attackers to train on discovered poisoned samples with flipped labels, producing counter-gradients that remove the backdoor without knowing the trigger. ZKP commitments lock the training data, which blocks adaptive attacks and gradient tampering. On CIFAR-10 under Dir(0.1) non-IID, it holds attack success rate below 6% against distributed backdoor and edge-case attacks.

TL;DR Proposes a ZKP and staking based federated learning defense that audits client samples and forces attackers to neutralize their own backdoors.
Open original ↗

Organisations named here

  • adaptive attack
  • CIFAR-10
  • Dir(0.1)
  • distributed backdoor attack
  • edge-case attack

Technologies named here

  • Zero-knowledge proofs
  • Staking mechanism
  • CIFAR-10

Techniques named here

  • Data poisoning in federated learning
  • Gradient tampering
  • Edge-case attacks, defended against on CIFAR-10 under Dir
  • Dynamic adaptive attacks based on reference model parameters

Related in AI Security