Search papers, labs, and topics across Lattice.
2
0
2
1
By separating logical reasoning from trajectory realization, DAG-STL enables zero-shot generalization in STL planning, outperforming direct diffusion-based approaches and even recovering performance comparable to model-based optimization with significantly reduced computation.
Offline reinforcement learning can achieve zero-shot generalization to unseen Signal Temporal Logic tasks by decoupling reachability learning from task-conditioned planning.
