Robust Parameter Learning for Uncertain MDPs

ArXi:2605.01339v1 Announce Type: new Learning-based approaches to verifying unknown Marko decision processes (MDPs) often employ uncertain MDPs. These models use, for example, confidence intervals to capture transition uncertainty and allow synthesis of policies that are robust to this uncertainty. However, this approach typically quantifies uncertainty independently for individual transition probabilities, ignoring dependencies due to shared latent quantities.