Test-Time Instance-Specific Parameter Composition: A New Paradigm for Adaptive Generative Modeling

ArXi:2603.27665v1 Announce Type: cross Existing generative models, such as diffusion and auto-regressive networks, are inherently static, relying on a fixed set of pretrained parameters to handle all inputs. In contrast, humans flexibly adapt their internal generative representations to each perceptual or imaginative context. Inspired by this capability, we