Search papers, labs, and topics across Lattice.
Shanghai Jiao Tong University, AI Lab
1
0
2
Sparse rewards can be transformed into actionable turn-level feedback, enabling agents to learn from both successful and misleading actions in long-horizon tasks.
