analiza aiqbench
This arXiv paper is about counterfactual reasoning in Markov Decision Processes, a method for explaining why an AI agent's sequential decisions led to failure and which actions would have changed the outcome. It contributes closed-form bounds and a robust intervention method that account for both estimation error in the MDP and ambiguity over which causal model is correct. For AI readers, the practical relevance is interpretability and failure analysis for decision-making systems, though the results are demonstrated on case studies and still need validation on realistic deployments.
Robust Counterfactuals for Markov Decision Processes under Statistical and Causal Uncertainties
arXiv:2610.20210v1 Announce Type: new Abstract: Counterfactual reasoning (asking what would have happened, had a decision been different) is a powerful tool for identifying causes of failure and explaining the behaviour of AI systems in sequential decision-making tasks. In Markov Decision Processes (MDPs), two sources of uncertainty complicate counterfactual inference: statistical uncertainty in the MDP transitions when the MDP is estimated from trajectory data, and uncertainty about the unknown causal model, as there are multiple structural causal models (SCMs) compatible with each MDP. Existing work addresses these two challenges in isolation. We address both simultaneously by introducing a framework for partial counterfactual inference in MDPs that is robust to both statistical and causal uncertainty. Our approach provides: 1) closed-form analytical bounds for the counterfactual probabilities that are valid across all SCMs entailed by a given interval MDP (IMDP); 2) a method to compute minimal robust interventions to achieve a desired counterfactual outcome via pessimistic value iteration on the induced counterfactual IMDP. Since the latter is constructed using our analytical bounds, our approach remains efficient and scalable to realistic MDP environments. Experiments across multiple case studies demonstrate our method's robustness and practical utility for identifying minimal sets of actions that, if changed, would have led to improved outcomes.
Abstract:Counterfactual reasoning (asking what would have happened, had a decision been different) is a powerful tool for identifying causes of failure and explaining the behaviour of AI systems in sequential decision-making tasks. In Markov Decision Processes (MDPs), two sources of uncertainty complicate counterfactual inference: statistical uncertainty in the MDP transitions when the MDP is estimated from trajectory data, and uncertainty about the unknown causal model, as there are multiple structural causal models (SCMs) compatible with each MDP. Existing work addresses these two challenges in isolation. We address both simultaneously by introducing a framework for partial counterfactual inference in MDPs that is robust to both statistical and causal uncertainty. Our approach provides: 1) closed-form analytical bounds for the counterfactual probabilities that are valid across all SCMs entailed by a given interval MDP (IMDP); 2) a method to compute minimal robust interventions to achieve a desired counterfactual outcome via pessimistic value iteration on the induced counterfactual IMDP. Since the latter is constructed using our analytical bounds, our approach remains efficient and scalable to realistic MDP environments. Experiments across multiple case studies demonstrate our method's robustness and practical utility for identifying minimal sets of actions that, if changed, would have led to improved outcomes.
Artificial Intelligence (cs.AI)
arXiv:2610.20210 [cs.AI]
(or arXiv:2610.20210v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.20210 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 295 cuvinte
Subiecte conexe
Robust Counterfactual Policy Optimisation via Nondeterministic Causal Models
arXiv · 5 aug. 2026
The Entangled Dimensions of Decision Theory
LessWrong · 30 iul. 2026
Runway’s WorldPrompt and the Engineering of Real-Time Worlds
Latent Space · 25 sept. 2026
CausAdv: A Causal-based Framework for Detecting Adversarial Examples
arXiv · 28 iul. 2026