Paper

{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.08578v1</id>\n    <title>Drift-to-Action Controllers: Budgeted Interventions with Online Risk Certificates</title>\n    <updated>2026-03-09T16:34:12Z</updated>\n    <link href='https://arxiv.org/abs/2603.08578v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.08578v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Deployed machine learning systems face distribution drift, yet most monitoring pipelines stop at alarms and leave the response underspecified under labeling, compute, and latency constraints. We introduce Drift2Act, a drift-to-action controller that treats monitoring as constrained decision-making with explicit safety. Drift2Act combines a sensing layer that maps unlabeled monitoring signals to a belief over drift types with an active risk certificate that queries a small set of delayed labels from a recent window to produce an anytime-valid upper bound $U_t(δ)$ on current risk. The certificate gates operation: if $U_t(δ) \\le τ$, the controller selects low-cost actions (e.g., recalibration or test-time adaptation); if $U_t(δ) &gt; τ$, it activates abstain/handoff and escalates to rollback or retraining under cooldowns. In a realistic streaming protocol with label delay and explicit intervention costs, Drift2Act achieves near-zero safety violations and fast recovery at moderate cost on WILDS Camelyon17, DomainNet, and a controlled synthetic drift stream, outperforming alarm-only monitoring, adapt-always adaptation, schedule-based retraining, selective prediction alone, and an ablation without certification. Overall, online risk certification enables reliable drift response and reframes monitoring as decision-making with safety.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CL'/>\n    <published>2026-03-09T16:34:12Z</published>\n    <arxiv:comment>Published as a conference paper at CAO Workshop at ICLR 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Ismail Lamaakal</name>\n    </author>\n    <author>\n      <name>Chaymae Yahyati</name>\n    </author>\n    <author>\n      <name>Khalid El Makkaoui</name>\n    </author>\n    <author>\n      <name>Ibrahim Ouahbi</name>\n    </author>\n    <author>\n      <name>Yassine Maleh</name>\n    </author>\n  </entry>"
}