Diagnosing and reducing exploitability in self-play tabular Q-learning using exact best responses, adversarial Bellman backups, and symmetry-aware state representations.
-
Updated
Sep 14, 2026 - Python
Diagnosing and reducing exploitability in self-play tabular Q-learning using exact best responses, adversarial Bellman backups, and symmetry-aware state representations.
Exact worst-case exploitability audits for language-model policies in imperfect-information games
REACHABLE AI SAST/SCA with reachability and exploitability proof; proposes reviewable fix PRs. Early access (beta).
Claude skill for CVE exploitability and mitigation intelligence.
To associate your repository with the exploitability topic, visit your repo's landing page and select "manage topics."