analiza aiqbench
This research introduces UnMaskFork (UMF), a novel framework applying test-time scaling techniques specifically to Masked Diffusion Language Models rather than standard autoregressive architectures. By utilizing Monte Carlo Tree Search and deterministic action branching, the method optimizes generation paths without relying on stochastic sampling, yielding significant performance gains in complex coding and mathematical reasoning tasks. The approach effectively bridges inference-time compute strategies between diffusion-based generative models and large language model benchmarks.
UnMaskFork: Test-Time Scaling for Masked Diffusion via Deterministic Action Branching
arXiv:2602.04344v2 Announce Type: replace-cross Abstract: Test-time scaling strategies have effectively leveraged inference-time compute to enhance the reasoning abilities of Autoregressive Large Language Models. In this work, we demonstrate that Masked Diffusion Language Models (MDLMs) are inherently amenable to advanced search strategies, owing to their iterative and non-autoregressive generation process. To leverage this, we propose UnMaskFork (UMF), a framework that formulates the unmasking trajectory as a search tree and employs Monte Carlo Tree Search to optimize the generation path. In contrast to standard scaling methods relying on stochastic sampling, UMF explores the search space through deterministic partial unmasking actions performed by multiple MDLMs. Our empirical evaluation demonstrates that UMF consistently outperforms existing test-time scaling baselines on complex coding benchmarks, while also exhibiting strong scalability on mathematical reasoning tasks.
Abstract:Test-time scaling strategies have effectively leveraged inference-time compute to enhance the reasoning abilities of Autoregressive Large Language Models. In this work, we demonstrate that Masked Diffusion Language Models (MDLMs) are inherently amenable to advanced search strategies, owing to their iterative and non-autoregressive generation process. To leverage this, we propose UnMaskFork (UMF), a framework that formulates the unmasking trajectory as a search tree and employs Monte Carlo Tree Search to optimize the generation path. In contrast to standard scaling methods relying on stochastic sampling, UMF explores the search space through deterministic partial unmasking actions performed by multiple MDLMs. Our empirical evaluation demonstrates that UMF consistently outperforms existing test-time scaling baselines on complex coding benchmarks, while also exhibiting strong scalability on mathematical reasoning tasks.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2602.04344 [cs.LG]
(or arXiv:2602.04344v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2602.04344 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 217 cuvinte
Subiecte conexe
UnMaskFork: Test-Time Scaling for Masked Diffusion via Deterministic Action Branching
Sakana AI · 21 iul. 2026
Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS · 4 sept. 2026
Frontier Risk Report (February to March 2026)
METR · 19 mai 2026
Informe de riesgos de la IA de frontera (febrero–marzo de 2026)
METR · 19 mai 2026