Reinforcement learning (RL) in robotics faces significant hurdles regarding sample efficiency and generalization across varying goals. While Offline RL mitigates the need for costly online interactions, its integration with goal-conditioned policies and transformer-based architectures remains underexplored. We introduce a Goal-Conditioned Decision Transformer adapted for offline multi-goal robotics. By explicitly incorporating goal states into the sequence modeling framework, our approach efficiently solves varying tasks using only pre-collected data. We validate this method on a newly released offline dataset for the Franka Emika Panda platform. Experimental results demonstrate that our approach outperforms state-of-the-art online baselines in complex tasks and maintains robustness in sparse-reward settings, even with limited expert demonstrations.
@article{arxiv.2410.06347,
title = {Goal-Conditioned Decision Transformer for Multi-Goal Offline Reinforcement Learning},
author = {Paweł Gajewski and Dominik Żurek and Marcin Pietroń and Kamil Faber},
journal= {arXiv preprint arXiv:2410.06347},
year = {2026}
}