Computationally lightweight classifiers with frequentist bounds on predictions

2026-03-23

AI Summary

This work addresses the challenge that high-accuracy classifiers often lack reliable uncertainty quantification, while kernel-based methods offering frequentist confidence bounds suffer from prohibitive $O(n^3)$ computational complexity, limiting their applicability in large-scale or resource-constrained safety-critical settings. The authors propose a lightweight classification algorithm based on Nadaraya–Watson kernel regression that, for the first time, delivers frequentist uncertainty intervals with dramatically reduced inference complexity—down to $O(n)$ or even $O(\log n)$. Experiments on synthetic data and the MIT-BIH electrocardiogram database demonstrate over 96% accuracy while providing uncertainty bounds capable of flagging low-confidence predictions. The method thus achieves a compelling balance of high accuracy, statistical reliability, and real-time efficiency, making it well-suited for embedded medical monitoring applications.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Uncertainty Representations
• Multiagent Systems: Multiagent Systems under Uncertainty

Application Category

• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

While both classical and neural network classifiers can achieve high accuracy, they fall short on offering uncertainty bounds on their predictions, making them unfit for safety-critical applications. Existing kernel-based classifiers that provide such bounds scale with $\mathcal O (n^{\sim3})$ in time, making them computationally intractable for large datasets. To address this, we propose a novel, computationally efficient classification algorithm based on the Nadaraya-Watson estimator, for whose estimates we derive frequentist uncertainty intervals. We evaluate our classifier on synthetically generated data and on electrocardiographic heartbeat signals from the MIT-BIH Arrhythmia database. We show that the method achieves competitive accuracy $>$\SI{96}{\percent} at $\mathcal O(n)$ and $\mathcal O(\log n)$ operations, while providing actionable uncertainty bounds. These bounds can, e.g., aid in flagging low-confidence predictions, making them suitable for real-time settings with resource constraints, such as diagnostic monitoring or implantable devices.

Problem

• uncertainty quantification
• computationally efficient classification
• frequentist bounds
• safety-critical applications
• kernel-based classifiers

Innovation

• computationally lightweight
• frequentist uncertainty bounds
• Nadaraya-Watson estimator
• real-time classification
• safety-critical applications