SIGNAL
Tracking the global AI frontier — labs · research · agents · policy
Frontier Signal
Research

Conformal Recovery-Deadline Certificates for Runtime Assurance of Adapting Controllers

Runtime assurance (RTA) protects a safety-critical system by switching from an advanced controller to a verified safe controller when a monitored condition is violated. The standard latching rule, which trips on the first breach of the safe set and then coasts, is correct for a diverging controller but pathological for a capable online-adapting one. Such a controller is unsafe by design during a bounded recovery transient. It must excite the plant to identify the fault before it can correct it,

Conformal Recovery-Deadline Certificates for Runtime Assurance of Adapting Controllers
Primary source tldr.takara.ai ↗

Published June 24, 2026 · Category: AI Research

Overview

Runtime assurance (RTA) protects a safety-critical system by switching from an advanced controller to a verified safe controller when a monitored condition is violated. The standard latching rule, which trips on the first breach of the safe set and then coasts, is correct for a diverging controller but pathological for a capable online-adapting one. Such a controller is unsafe by design during a bounded recovery transient. It must excite the plant to identify the fault before it can correct it, so a latching shield trips on that transient and suppresses a controller that would have recovered. We introduce the conformal recovery-deadline certificate, a split-conformal, distribution-free, finite-sample upper bound on the adapting controller's recovery time that licenses delayed fallback with a coverage guarantee, backstopped by a verified monitor at a hard critical limit. The certified deadline discriminates capable from incapable controllers, keeping the recoverer autonomous while catching the diverger. The construction separates autonomy, governed by statistical coverage, from safety, governed by the verified backstop, as an instance of reliability-asymmetric design. We prove marginal coverage, a weighted extension that restores coverage under a known fault-distribution shift, and group-conditional Mondrian coverage. We demonstrate all three on two unrelated Simplex testbeds: a 6-DOF spacecraft attitude controller and a torque-controlled inverted pendulum. Both show the same suppression pathology and the same cure, making the certificate a domain-general mechanism rather than a single-system trick.

Source

Originally published at tldr.takara.ai.

Related Articles

F
Frontier Signal Desk

Frontier Signal tracks the global AI frontier — labs, research, agents, creation tools and real-world practice — straight from primary sources. Tip the desk: editorial@news.tunx.ai

Email the desk →
From our network: explore the AI assistant platform behind this site. Visit tunx.ai →
Note: This story is aggregated and summarized from the primary source linked above; the original publisher retains all rights. Details may evolve after publication — always confirm against the source. Nothing here is professional, legal or investment advice.

Related Stories

More from Research →