Diagnosing Non-Markovian Observations in Reinforcement Learning via Prediction-Based Violation Scoring
Provides a diagnostic tool for practitioners to detect and address non-Markovian observations in RL, which is a practical problem but the method is incremental (combining random forest and ridge regression).
The paper introduces a prediction-based scoring method to quantify non-Markovian structure in observation trajectories for RL, achieving significant positive monotonicity between noise intensity and violation score in 7 of 16 environment-algorithm pairs (Spearman rho up to 0.78) and demonstrating that the score can guide architecture selection to recover performance lost to non-Markovian observations.
Reinforcement learning algorithms assume that observations satisfy the Markov property, yet real-world sensors frequently violate this assumption through correlated noise, latency, or partial observability. Standard performance metrics conflate Markov breakdowns with other sources of suboptimality, leaving practitioners without diagnostic tools for such violations. This paper introduces a prediction-based scoring method that quantifies non-Markovian structure in observation trajectories. A random forest first removes nonlinear Markov-compliant dynamics; ridge regression then tests whether historical observations reduce prediction error on the residuals beyond what the current observation provides. The resulting score is bounded in [0, 1] and requires no causal graph construction. Evaluation spans six environments (CartPole, Pendulum, Acrobot, HalfCheetah, Hopper, Walker2d), three algorithms (PPO, A2C, SAC), controlled AR(1) noise at six intensity levels, and 10 seeds per condition. In post-hoc detection, 7 of 16 environment-algorithm pairs, primarily high-dimensional locomotion tasks, show significant positive monotonicity between noise intensity and the violation score (Spearman rho up to 0.78, confirmed under repeated-measures analysis); under training-time noise, 13 of 16 pairs exhibit statistically significant reward degradation. An inversion phenomenon is documented in low-dimensional environments where the random forest absorbs the noise signal, causing the score to decrease as true violations grow, a failure mode analyzed in detail. A practical utility experiment demonstrates that the proposed score correctly identifies partial observability and guides architecture selection, fully recovering performance lost to non-Markovian observations. Source code to reproduce all results is provided at https://github.com/NAVEENMN/Markovianes.