Interventional Causal Circuits for Safe Robot Action Testing and Failure Recovery

2026-07-16

AI Summary

This work addresses the inefficiency of robotic action safety testing, which often suffers from high computational costs, blind resampling, and a lack of convergence guarantees. The paper proposes the first closed-loop testing framework that integrates causal intervention reasoning by combining joint probability trees (JPTs) with causal circuits based on marginal certainty variable trees. Without requiring retraining or additional data, the method enables precise causal diagnosis of rejected actions, generates interpretable reports, and automatically filters out out-of-distribution candidates. It supports exact intervention queries in polynomial time. In ROS2 simulations, the approach reduces failed attempts by 10.3% under high-quality JPTs and by 37% under degraded JPTs, while producing structured causal reports for every rejection to facilitate autonomous recovery.

Abstract

Safe physical AI for robot actions are required not only likely to succeed but tested to be safe before execution. In practice, however, formal testing of motion parameters is computationally expensive, and the cost scales poorly with the dimensionality of the action space. When a proposed action is rejected by a tester, the naive response is to resample blindly until a passing candidate is found. This is wasteful, uninformative, and offers no convergence. We argue that rejection should instead trigger causal diagnosis: a principled identification of which action parameter caused the failure and what corrective value maximises the probability of passing testing under the interventional probability distribution. We propose a closed-loop framework that couples a Joint Probability Tree (JPT) with a Causal Circuit derived from a Marginal-Deterministic Variable Tree, enabling exact polytime computation without retraining, or additional data collection. The framework validates tractability of all interventional queries before the robot begins operating, and out-of-support candidates are detected and excluded from correction automatically. We perform experiments in a ROS2 simulation environment, and the framework demonstrates complementary roles across quality of distribution: under a high-quality JPT, the Causal Circuit reduces failed attempts by 10.3% and under a degraded JPT, it reduces total failed attempts by 37%. Every rejected plan produces a structured, interpretable causal report naming the primary cause variable, its observed value, and the recommended corrective region, supporting operator oversight and autonomous recovery without a separately trained failure model.

Problem

• safe robot action testing
• interventional causal inference
• failure recovery
• action parameter diagnosis
• computational tractability

Innovation

• Causal Circuit
• Interventional Probability
• Joint Probability Tree
• Safe Robot Action
• Failure Recovery