Skip to content
NewA calmer, faster ZeroHour

Paper proposes revision-bound evidence cuts for authenticated queries

Original title: Revision-Bound Evidence Cuts: Authenticated Queries and Contextual Witness Capacity

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

A verifier checked 16,824 records, and the SCENE prototype scaled signed revisions to one million records.

Researchers present revision-bound evidence cuts for authenticating supporting and opposing inference obligations as source versions change. An authenticated relational query contract binds source records, selected bags, counting units, parameters, and aggregates independently of execution strategy. A verifier checked 16,824 Chogha Mish records, while a 340-row NISP cohort showed a weighting-rule change can reverse a comparison. The SCENE prototype evaluated signed revisions and adversarial mutations at up to one million records. The authors say the guarantees authenticate specified computations, not historical truth or institutional endorsement.

Unconfirmed: The paper does not report independent reproduction of the prototype results.

Key facts
  • Three valid query strategies agreed, and eight contract or view mutations were rejected.
  • Changing an admitted weighting rule reversed a comparison while preserving every source record.
  • SCENE evaluated scoped disclosures and adversarial mutations at one million source records.
  • Guarantees depend on declared policies and independently pinned trust roots.
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

Authentic records can support unauthorized narratives when their qualifications, query scope, or counterevidence are silently dropped. We present revision-bound evidence cuts, a framework for authenticating both supporting and opposing inference obligations across changing source versions.

arXiv cs.CR

By Faruk Alpay · 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.