AuditFL: A Contract-Backed Clearing Layer for Open Federated Learning
AuditFL adds on-chain auditability to federated learning using contract-backed tickets and off-chain model storage
The paper presents AuditFL, a clearing layer for open federated learning that keeps training off-chain while recording identifiers and economic state on-chain. Clients reserve update tickets with deposits for auditable rounds. It was implemented as a state machine and Solidity contract on Anvil. On MNIST with up to 200 clients, it preserves learning trajectory with linear overhead under stress tests.
- Deterministic settlement rules for checkpoints
- 5,026 transactions at 200 clients
- 2,413 off-chain artifacts stored
- Handles controlled dropout and invalid updates
From the original article
1 min read at huggingface.co
Federated learning (FL) keeps raw data local, but most deployments still rely on a trusted coordinator for client admission, checkpoint distribution, update collection, aggregation, and rewards. This paper presents AuditFL, a contract-backed clearing layer for auditable FL in open settings.
By Alexandros Balla, Spyros Lalis · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.