Single-Ballot Risk-Limiting Audits Using Convex Optimization

By Stephen Checkoway, Anand Sarwate, and Hovav Shacham.

In Proceedings of EVT/WOTE 2010. USENIX/ACCURATE/IAVoSS, August 2010.

Abstract

We take an information-theoretic approach to sequential election auditing. By comparing how far an empirical distribution of audited votes diverges from any distribution in which the reported outcome is incorrect, we gain a high degree of confidence in the outcome when our procedure confirms the reported results.

Material

• Proceedings version in PDF.
• Slides from EVT/WOTE in PDF.
• Full version, local copy in PDF, PS, and DVI.

Reference

@InProceedings{checkoway-sarwate-shacham:single-ballot:evt10,
  author =       {Stephen Checkoway and Anand Sarwate and Hovav Shacham},
  title =        {Single-Ballot Risk-Limiting Audits Using Convex Optimization},
  booktitle =    {Proceedings of EVT/WOTE 2010},
  editor =       {Doug Jones and Jean-Jacques Quisquater and Eric Rescorla},
  organization = {USENIX/ACCURATE/IAVoSS}
  year =         2010,
  month =        aug,
  url =          {https://checkoway.net/papers/audits2010},
}