[S31] Sep 28: Discrete Beckmann Transport Models for One-Step Language Modeling and Reasoning

[Diffusion LLM]

Hello folks,

This Monday, Sophia Tang will present her recent paper:

Paper: Discrete Beckmann Transport Models for One-Step Language Modeling and Reasoning

Meeting Link: click here

Time: Sep 28 (Monday) 1pm ET / 10am PT / 7pm CET / 10:30pm IST

Abstract: Discrete diffusion and flow models are a promising alternative to autoregressive language models, but compressing many-step sampling into fewer steps typically requires distilling a pretrained teacher model. This caps the student at the teacher's quality and requires a costly two-stage training pipeline. We introduce Discrete Beckmann Transport Models (DBTM), built on a time-independent flow whose autonomous transport map provably carries any point in the ambient space to a fixed point on the vertices of the simplex in a single step. We show that this fixed-point property is characterized by a conservation equation whose residual can be minimized directly from data, removing the requirement for a teacher flow and time conditioning. Under this construction, a partially trained map corresponds to the flow truncated at finite time, so generation reduces to iterating one map until it reaches a fixed point. We further extend the map to a partial-context interpolant where additional function evaluations act as refinement steps rather than ODE integration steps. On language modeling and reasoning tasks, DBTM enables one- and few-step generation that improves quality and accuracy over discrete diffusion and continuous flow baselines