Reproduce “Defensive Sufficiency in a Stackelberg Model of AI Security” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Subhabrata Majumdar, Rajlakshmi Chavan

The authors analyze repeated vulnerability discovery, repair, and defenders’ investment choices. Given discovery, effective-repair, and retention assumptions, a finite attack surface is eventually covered with probability one. In an illustrative comparison, joint investment yields compromise probability 0.58 and expected compromise time 0.25, versus 0.58 probability and 0.58 time units with fixed repair; these model results do not establish deployed-system performance.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.