AdaptNC: Adaptive Nonconformity Scores for Conformal Prediction under Distribution Shift

TL;DR

AdaptNC introduces an adaptive framework for conformal prediction that jointly adjusts nonconformity scores and thresholds, improving uncertainty quantification under distribution shifts in robotics.

Contribution

It proposes a novel joint adaptation method for nonconformity scores and thresholds, enhancing prediction efficiency and stability during environment changes.

Findings

AdaptNC reduces prediction region volume significantly compared to baselines.

It maintains target coverage levels despite distribution shifts.

The method is effective across diverse robotic benchmarks.

Abstract

Rigorous uncertainty quantification is essential for the safe deployment of autonomous systems in unconstrained environments. Conformal Prediction (CP) provides a distribution-free framework for this task, yet its standard formulations rely on exchangeability assumptions that are violated by the distribution shifts inherent in real-world robotics. Existing online CP methods maintain target coverage by adaptively scaling the conformal threshold, but typically employ a static nonconformity score function. We show that this fixed geometry leads to highly conservative, volume-inefficient prediction regions when environments undergo structural shifts. To address this, we propose $\textbf{AdaptNC}$, a framework for the joint online adaptation of both the nonconformity score parameters and the conformal threshold. AdaptNC leverages an adaptive reweighting scheme to optimize score functions, and introduces a replay buffer mechanism to mitigate the coverage instability that occurs during score transitions. We evaluate AdaptNC on diverse robotic benchmarks involving multi-agent policy changes, environmental changes and sensor degradation. Our results demonstrate that AdaptNC significantly reduces prediction region volume compared to state-of-the-art threshold-only baselines while maintaining target coverage levels.