Walker-2D
Emerging4papers using it
2025first seen
The 'Walker2D' dataset/benchmark is a simulated environment used to evaluate model-based reinforcement learning algorithms, specifically focusing on the performance of agents in learning to walk in a two-dimensional space.
Papers using Walker-2D (4)
- ARISE: Adaptive Reinforcement Integrated with Swarm ExplorationOvercoming Overfitting in Reinforcement Learning via Gaussian Process Diffusion PolicyDiagnosing Non-Markovian Observations in Reinforcement Learning via Prediction-Based Violation ScoringAdvantage-Guided Diffusion for Model-Based Reinforcement Learning