Search papers, labs, and topics across Lattice.
University of California
2
0
3
Boundary singularities usually cause learned optimizers to fail on constrained problems, but coupling primal-dual shifts with learned diagonal preconditioners slashes exact interior-point solver iterations by up to 67% without ever computing a Hessian.
By fusing orthogonalized momentum with adaptive noise scaling, NAMO and NAMO-D offer a surprisingly simple recipe for faster and more stable LLM training compared to AdamW and Muon.
