Variational Learning for Insertion-based Generation

ArXi:2606.02133v1 Announce Type: cross Non-monotonic sequence generation methods, such as masked diffusion models, provide a flexible alternative to left-to-right autoregressive modeling by allowing tokens to be generated in non-fixed and prescribed orders. Despite their practical advantages, most existing non-monotonic models are order-agnostic and rely on a fixed-length grid, limiting their ability to variable-length generation and adaptive insertion order. In this work, we