Measuring Prediction Uncertainty in Neural Cellular Automata

2026-05-26

AI Summary

This work addresses the lack of reliable predictive uncertainty quantification in neural cellular automata (NCAs) for medical image segmentation, which hinders the assessment of output trustworthiness. Treating NCAs as dynamical systems, the study introduces a novel metric termed “resilience,” which probes the stability of attractors through infinitesimal perturbations to internal states. This approach enables non-intrusive uncertainty estimation without requiring architectural modifications or retraining. By integrating perturbation analysis, dynamical systems theory, and selective prediction evaluation, the proposed method consistently outperforms existing baselines across multiple medical segmentation benchmarks. It demonstrates a superior ability to identify failure cases, thereby enhancing model reliability and safety in clinical applications.

Abstract

Neural cellular automata (NCA) provide a lightweight alternative to encoder-decoder segmentation networks. However, it can be difficult to decide when a prediction should be trusted. Here, we study uncertainty estimation for NCA-based medical image segmentation without modifying the underlying architecture or retraining the model. Our approach is motivated by viewing the NCA as a dynamical system where convergent attractors correspond to confident predictions. Concretely, we propose resilience, a simple measure that leverages the intrinsic iterative structure of NCAs by probing the stability of the final prediction under small perturbations of the automaton state. Predictions that return to the same solution are deemed confident, while those that change substantially are flagged as uncertain. We evaluate uncertainty by its ability to predict segmentation quality using selective prediction metrics ($Δ$Dice@90 and AURC) and ranking metrics (AUROC and AUPRC). Across multiple medical segmentation benchmarks, resilience identifies failure cases more reliably than baselines, improving trust and safety in NCA-based models.

Problem

• Neural Cellular Automata
• Uncertainty Estimation
• Medical Image Segmentation
• Prediction Confidence
• Selective Prediction

Innovation

• Neural Cellular Automata
• Uncertainty Estimation
• Resilience
• Medical Image Segmentation
• Selective Prediction