Skip to content
NewA calmer, faster ZeroHour

Paper states exact resource laws for hidden integer feedback

Original title: Exact Compilation and Resource Laws for Hidden Integer Feedback

arXiv cs.CRby Ganxuan Yang ()1 min read1 · 1 read
AI summary · x-ai/grok-4.7Summaries by Vani AI

Keeping carries until an observed echo yields exact compilation with support-width exponent 3C-2 for budget C at least 2.

An arXiv paper studies a rational-feedback mechanism in which hidden randomized rounding couples privacy loss across successive integer releases, under predictable signed-unit participation and full observation of physical outputs. Preserving complete carries until an observed echo gives exact finite-law point compilation. Releasing commitments earlier produces a monotone hierarchy of calibration loss, and legal regeneration yields an unrestricted KL rate with a square-root testing remainder. Finite-window programs approximate the asymptotic frontier, and a matched static workload compares a structured integer matrix with scalar feedback.

Key facts
  • Support-width exponent is 3C-2 at every fixed budget C of at least 2.
  • A loss allowance of 0.09 at one discrete Gaussian requires all four commitments.
  • Windowed occupation programs reach frontier error 2 to the minus b in polynomial bit time.
  • The matrix design improves error and fresh randomness versus scalar feedback.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

1 min read at arxiv.org

Hidden randomized rounding couples the privacy loss of successive integer releases. We study a rational-feedback mechanism under predictable signed-unit participation and full observation of every physical output. A preparation's lifetime connects three results.

arXiv cs.CR

By Ganxuan Yang · arxiv.org

Read the full article at arXiv cs.CR

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.