Search papers, labs, and topics across Lattice.
2
0
3
4
This work proposes MATCH, a closed-loop framework for model-aware tool learning with curriculum scheduling and hierarchically gated rewards, and shows consistent improvements across four backbones from two model families.
MagicSelector achieves unprecedented tool retrieval accuracy by translating vague user instructions into precise subtasks, outperforming state-of-the-art methods.
