M\(^3\)PC: Test-time Model Predictive Control For Pretrained Masked Trajectory Model

Abstract

Recent work in Offline Reinforcement Learning (RL) has shown that a unified Transformer trained under a masked auto-encoding objective can effectively capture the relationships between different modalities (e.g., states, actions, rewards) within given trajectory datasets. However, this information has not been fully exploited during the inference phase, where the agent needs to generate an optimal policy instead of just reconstructing masked components from unmasked ones. Given that a pretrained trajectory model can act as both a Policy Model and a World Model with appropriate mask patterns, we propose using Model Predictive Control (MPC) at test time to leverage the model's own predictive capability to guide its action selection. Empirical results on D4RL and RoboMimic show that our inference-phase MPC significantly improves the decision-making performance of a pretrained trajectory model without any additional parameter training. Furthermore, our framework can be adapted to Offline t

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keywen2024m

Related papers

• Reprem: Representation Pre-training With Masked Model For Reinforcement Learning (2023)0.00
• Model Predictive Control With Self-supervised Representation Learning (2023)0.00
• Masked Autoencoding For Scalable And Generalizable Decision Making (2022)0.00
• Decision Mamba: A Multi-grained State Space Model With Self-evolution Regularization For Offline RL (2024)0.00
• Policy Search Using Dynamic Mirror Descent MPC For Model Free Off Policy RL (2021)0.00
• Adaptive Probabilistic Trajectory Optimization Via Efficient Approximate Inference (2016)0.00
• TD-MPC2: Scalable, Robust World Models For Continuous Control (2023)0.00
• Offline Trajectory Optimization For Offline Reinforcement Learning (2024)1.20