mlsanityData to 14 Oct 2026

T^star: Progressive Block Scaling for Masked Diffusion Language Models Through Trajectory Aware Reinforcement Learning

arXiv 2601.11214 ↗First posted 16 Jan 2026cs.CLcited by 5

On the map: Diffusion models → Discrete and text diffusion

Abstract (arXiv)

We present T^star, a simple TraceRL-based training curriculum for progressive block-size scaling in masked diffusion language models (MDMs). Starting from an AR-initialized small-block MDM, T^star transitions smoothly to larger blocks, enabling higher-parallelism decoding with minimal performance degradation on math reasoning benchmarks. Moreover, further analysis suggests that T^star may actually converge to an alternative decoding schedule that achieves comparable performance.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 9 of its 17 references that are in the map. First posted on arXiv 16 Jan 2026.

Builds on

9 of 17 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025SDAR: A Synergistic Diffusion-AutoRegression Paradigm for Scalable Sequence Generation7 Oct
  2. 2025DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation25 Jun
  3. 2024Simple and Effective Masked Diffusion Language Models11 Jun

Other references · 6

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  6. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 2,026

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026DreamReasoner-8B: Block-Size Curriculum Learning for Diffusion Reasoning Modelscited by 5
  2. 2026BARD: Bridging AutoRegressive and Diffusion Vision-Language Models Via Highly Efficient Progressive Block Merging and Stage-Wise Distillationcited by 2
  3. 2026ForkLeft: Entropy-First Rollouts for Prefix-Aligned Autoregressive-to-Diffusion Distillationcited by 1
  4. 2026SemBlock: Semantic Boundary Dynamic Blocks for Diffusion LLMscited by 0
  5. 2026d-OPD: Future-Aware On-Policy Distillation for Block Diffusion Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026TRIMS: Trajectory-Ranked Instruction Masked Supervision for Diffusion Language Models
  2. 2026Mask-Aware Policy Gradients for Diffusion Language Models
  3. 2025Taming Masked Diffusion Language Models via Consistency Trajectory Reinforcement Learning with Fewer Decoding Step
  4. 2026SLIM-RL: Risk-Budgeted Random-Masking RL for Diffusion LLMs Without Trajectory Slicing
  5. 2025dUltra: Ultra-Fast Diffusion Language Models via Reinforcement Learning
  6. 2026Read the Trace, Steer the Path: Trajectory-Aware Reinforcement Learning for Diffusion Language Models
  7. 2026UnMaskFork: Test-Time Scaling for Masked Diffusion via Deterministic Action Branching
  8. 2026x-Prediction Flow: Efficient Continuous Decoding for Masked Diffusion Language Models