Discrete Beckmann Transport Models for One-Step Language Modeling and Reasoning
- Published
- Source
- arXiv
- Paper number
- 1088
- Field
- Machine Learning
- arXiv ID
- 2609.15903
Key points
- The paper establishes theory showing that a one-step transport map satisfies the conservation equation b·∇T=0, and builds a method that trains directly by minimizing its residual without teacher distillation.
- Iteratively applying the partially learned map converges to a fixed point (the completed sentence), so accuracy can be raised while reducing the number of computation steps.
- The method supports self-correcting generation that uses extra computation as a refinement step via partial-context interpolation.
- On language modeling (OpenWebText generation PPL) and reasoning tasks, it outperformed diffusion and flow-map baselines such as FMLM with equal or fewer function evaluations.
Paper links
External research summaries. These are not HDATF publications or measured product results.