Search papers, labs, and topics across Lattice.
1
0
3
1
Ditch the costly process supervision: EVOM trains LLMs to generate optimization code by treating solvers as deterministic reward functions, enabling zero-shot solver transfer and efficient adaptation.
