TRIAGE: Type-Routed Interventions via Aleatoric-Epistemic Gated Estimation in Robotic Manipulation and Adaptive Perception -- Don't Treat All Uncertainty the Same

2026-03-09

AI Summary

This work addresses a critical limitation in existing uncertainty-aware robotic systems, which collapse uncertainty into a single scalar and fail to distinguish between aleatoric (observational noise) and epistemic (model mismatch) uncertainties, leading to suboptimal responses. To overcome this, the authors propose a lightweight post-processing framework that orthogonally decomposes these two uncertainty types during inference: aleatoric uncertainty is estimated via a Mahalanobis distance–based density model, while epistemic uncertainty is detected using a noise-robust ensemble of forward dynamics models. A gating mechanism then triggers targeted interventions—high aleatoric uncertainty prompts observation recovery, whereas high epistemic uncertainty modulates control actions and guides model selection in adaptive perception. Evaluated in closed-loop execution, this approach achieves type-aware uncertainty handling, boosting task success rates from 59.4% to 80.4% (+21.0%) under composite perturbations and reducing computational cost by 58.2% on the MOT17 tracking benchmark with only a 0.4% accuracy drop.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Most uncertainty-aware robotic systems collapse prediction uncertainty into a single scalar score and use it to trigger uniform corrective responses. This aggregation obscures whether uncertainty arises from corrupted observations or from mismatch between the learned model and the true system dynamics. As a result, corrective actions may be applied to the wrong component of the closed loop, degrading performance relative to leaving the policy unchanged. We introduce a lightweight post hoc framework that decomposes uncertainty into aleatoric and epistemic components and uses these signals to regulate system responses at inference time. Aleatoric uncertainty is estimated from deviations in the observation distribution using a Mahalanobis density model, while epistemic uncertainty is detected using a noise robust forward dynamics ensemble that isolates model mismatch from measurement corruption. The two signals remain empirically near orthogonal during closed loop execution and enable type specific responses. High aleatoric uncertainty triggers observation recovery, while high epistemic uncertainty moderates control actions. The same signals also regulate adaptive perception by guiding model capacity selection during tracking inference. Experiments demonstrate consistent improvements across both control and perception tasks. In robotic manipulation, the decomposed controller improves task success from 59.4% to 80.4% under compound perturbations and outperforms a combined uncertainty baseline by up to 21.0%. In adaptive tracking inference on MOT17, uncertainty-guided model selection reduces average compute by 58.2% relative to a fixed high capacity detector while preserving detection quality within 0.4%. Code and demo videos are available at https://divake.github.io/uncertainty-decomposition/.

Problem

• uncertainty decomposition
• aleatoric uncertainty
• epistemic uncertainty
• robotic manipulation
• adaptive perception

Innovation

• uncertainty decomposition
• aleatoric uncertainty
• epistemic uncertainty
• adaptive perception
• robotic manipulation