The paper describes an algorithm that uses an integral score function to quantify the lumped effect of disturbance and learning error. Mechanism: The integral score aggregates disturbance and learning error into an online uncertainty estimate; authors report long-run coverage guarantees and synthesis with safety-critical controllers such as robust tube MPC. The arXiv record says the paper was accepted to CDC 2026.
Why now: The result addresses a practical gap in safety-critical adaptive control: conventional online conformal prediction is insufficient when systems adapt their dynamics without observations of state derivatives. The paper’s numerical demonstration uses an all-layer adaptive deep neural network controlling a quadcopter with robust tube MPC. First implication: Researchers can evaluate SI-OCP as an uncertainty-estimation component in adaptive control stacks.
Evidence boundary: The supplied arXiv record describes a numerical simulation and reports the authors’ guarantees; it does not establish real-world flight testing, deployment, or performance against other methods. The supplied material does not substantiate repository contents or reproducibility claims.
First watch condition: Look for independently inspectable code, benchmark comparisons, or physical flight tests that clarify how the method behaves beyond the reported simulation. Acceptance and a simulation make this a distinct research record; no prior New Runtime record was supplied as a basis for linking it to an existing analytical thread.