Scale-Wise VAR is Secretly Discrete Diffusion

ICLR 2026 Conference Submission14309 Authors

Keywords: Visual Autoregressive Generation, Discrete Diffusion

Abstract: Autoregressive (AR) transformers have emerged as a powerful paradigm for visual generation, largely due to their scalability, computational efficiency and unified architecture with language and vision. Among them, next scale prediction Visual Autoregressive Generation (VAR) has recently demonstrated remarkable performance, even surpassing diffusion-based models. In this work, we revisit VAR and uncover a theoretical insight: when equipped with a Markovian attention mask, VAR is mathematically equivalent to a discrete diffusion. We term this reinterpretation as Scalable Visual Refinement with Discrete Diffusion (SRDD), establishing a principled bridge between AR transformers and diffusion models. Leveraging this new perspective, we show how one can directly import the advantages of diffusion—such as iterative refinement and and reduce architectural inefficiencies into VAR, yielding faster convergence, lower inference cost, and improved zero-shot reconstruction. Across multiple datasets, we show that the diffusion-based perspective of VAR leads to consistent gains in efficiency and generation. To facilitate further research, we will make the code and models public.

Primary Area: interpretability and explainable AI

Submission Number: 14309