Bifurcation Models: Learning Set-Valued Solution Maps with Weight-Tied Dynamics
📰 ArXiv cs.AI
arXiv:2605.07277v1 Announce Type: cross Abstract: Many scientific and combinatorial problems admit multiple correct solutions, not a single label. Standard supervised learning resolves this ambiguity by choosing one solution as the target, but this hidden selector can be arbitrary, discontinuous, and harder to learn than the underlying solution set. We study bifurcation models, a weight-tied dynamical view in which different initializations can converge to different stable equilibria, so the mod
Full Article
Title: Bifurcation Models: Learning Set-Valued Solution Maps with Weight-Tied Dynamics
Abstract:
arXiv:2605.07277v1 Announce Type: cross Abstract: Many scientific and combinatorial problems admit multiple correct solutions, not a single label. Standard supervised learning resolves this ambiguity by choosing one solution as the target, but this hidden selector can be arbitrary, discontinuous, and harder to learn than the underlying solution set. We study bifurcation models, a weight-tied dynamical view in which different initializations can converge to different stable equilibria, so the mod
Abstract:
arXiv:2605.07277v1 Announce Type: cross Abstract: Many scientific and combinatorial problems admit multiple correct solutions, not a single label. Standard supervised learning resolves this ambiguity by choosing one solution as the target, but this hidden selector can be arbitrary, discontinuous, and harder to learn than the underlying solution set. We study bifurcation models, a weight-tied dynamical view in which different initializations can converge to different stable equilibria, so the mod
DeepCamp AI