Search papers, labs, and topics across Lattice.
Duke University
2
0
2
Energy-guided diffusion models can bridge the gap between source and target domains in off-dynamics offline RL, enabling effective trajectory generation without retraining the diffusion model.
By clustering transitions and estimating cluster-level dynamics discrepancy, LoDADA offers a fine-grained and scalable data selection strategy that outperforms existing off-dynamics offline RL methods.
