{"id":"arxiv/2610.31361","url":"https://arxiv.org/abs/2610.31361","tags":["math.OC","math.PR","math.ST"],"score":0,"title":"Gradient Discrete Schrodinger Bridge Matching for Peptide Generation","domain":"math.OC","source":"arxiv","authors":["Abdallah BenAbdallah","Mohsen Dlala"],"reviews":0,"subject":"math.OC","summary":"The optimal control of a discrete Schrödinger bridge is a discrete gradient, a difference of vertex potentials, but recent models such as MadSBM instead learn a free control on the edges by cross-entropy. We prove that this leaves the action uncontrolled: cross-entropy identifies the control only up to a state-dependent shift (a gauge), and the action is $\\operatorname{KL}(r\\|p)$ plus a gauge term the loss never constrains. We propose GDSBM, which writes the control as a difference of vertex potentials and adds the action to the loss; pinning the gauge costs no expressiveness and makes the…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-10-22","reactions":{}}