Newest papers

Beyond the Sequentiality Bottleneck: A Hybrid Diffusion-Autoregressive Framework for Efficient Multimodal Reasoning

Wei Cai, Jian Zhao, Yuchen Yuan, Xuelong Li · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28584 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

Los modelos autoregresivos (ARMs) se enfrentan a un "bottleneck de secuencia" en el razonamiento, mientras que los modelos de lenguaje de difusión (DLMs) ofrecen paralelismo pero carecen de coherencia semántica. Para abordar ambas limitaciones, proponemos Collaborative Reasoning Asymmetric Learning (CoRAL), un marco híbrido a través de una arquitectura Asymmetric Planner-Executor (APE): un DLM Planner genera K diversas trayectorias de razonamiento en paralelo, mientras que un gran ARM Executor selecciona el plan óptimo y sintetiza la respuesta final. Para entrenar al Planner de manera efectiva, introducimos Reward-Guided Denoising for CoT (RDC), que integra recompensas basadas en resultados en la difusión para

Machine translation (Meta M2M100); the English original follows.

Abstract

Autoregressive models (ARMs) face a "Sequentiality Bottleneck" in reasoning, while diffusion language models (DLMs) offer parallelism but lack semantic coherence. To addresses both limitations, we propose Collaborative Reasoning Asymmetric Learning (CoRAL), a hybrid framework via an Asymmetric Planner-Executor (APE) architecture: a DLM Planner generates K diverse reasoning trajectories in parallel, while a large ARM Executor selects the optimal plan and synthesizes the final answer. To train the Planner effectively, we introduce Reward-Guided Denoising for CoT (RDC), which integrates outcome-based rewards into the diffusion process to steer it toward goal-directed reasoning. Experiments on vision-language benchmarks show that CoRAL significantly outperforms prior diffusion-based methods and achieves competitive performance against strong ARMs, achieving both superior efficiency and quality.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Artificial intelligence, Planner, Process (computing), Face (sociological concept), Plan (archaeology)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange