Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching

• URL: http://arxiv.org/abs/2604.15009v1
• Date: Thu, 16 Apr 2026 13:36:04 GMT
• Title: Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching
• Abstract summary: We develop a non-autoregressive (NAR) language modeling approach, named YAN, instantiated with both Transformer and Mamba architectures.<n>Across multiple downstream tasks, YAN achieves generation quality on par with both autoregressive (AR) and diffusion-based NAR language models, while requiring as few as three sampling steps.<n>This yields a $40times$ speedup over AR baselines and up to a $103times$ speedup over diffusion language models, demonstrating substantial efficiency advantages for language modeling.
• Score: 1.4653929698362125
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Flow matching retains the generation quality of diffusion models while enabling substantially faster inference, making it a compelling paradigm for generative modeling. However, when applied to language modeling, it exhibits fundamental limitations in representing complex latent distributions with irregular geometries, such as anisotropy and multimodality. To address these challenges, we propose a mixture-of-experts flow matching (MoE-FM) framework, which captures complex global transport geometries in latent space by decomposing them into locally specialized vector fields. Building on MoE-FM, we develop a non-autoregressive (NAR) language modeling approach, named YAN, instantiated with both Transformer and Mamba architectures. Across multiple downstream tasks, YAN achieves generation quality on par with both autoregressive (AR) and diffusion-based NAR language models, while requiring as few as three sampling steps. This yields a $40\times$ speedup over AR baselines and up to a $10^3\times$ speedup over diffusion language models, demonstrating substantial efficiency advantages for language modeling.

Related papers

• Omni-Diffusion: Unified Multimodal Understanding and Generation with Masked Discrete Diffusion [66.78268790956493]
  We introduce Omni-Diffusion, the first any-to-any multimodal language model built entirely on mask-based discrete diffusion models.<n>Our method outperforms or performs on par with existing multimodal systems that process two or more modalities.
  arXiv  Detail & Related papers  (2026-03-06T18:59:57Z)
• One-step Language Modeling via Continuous Denoising [36.18484491074519]
  We show that language models leveraging flow-based continuous denoising can outperform discrete diffusion in both quality and speed.<n>Our work calls into question the widely held hypothesis that discrete diffusion processes are necessary for generative modeling over discrete modalities.
  arXiv  Detail & Related papers  (2026-02-18T19:23:07Z)
• Autoregressive Models Rival Diffusion Models at ANY-ORDER Generation [35.63237650402896]
  We propose Any-order Any-subset Autoregressive modeling (A3)<n>A3 is a framework that extends the standard AR factorization to arbitrary token groups and generation orders.<n> Experiments on question answering, commonsense reasoning, and story infilling demonstrate that A3 outperforms diffusion-based models.
  arXiv  Detail & Related papers  (2026-01-19T17:03:48Z)
• Towards Latent Diffusion Suitable For Text [7.293508593001522]
  We introduce Neural Flow Diffusion Models for language generation, an extension of NFDM that enables the straightforward application of continuous diffusion models to discrete state spaces.<n>Our model substantially reduces the likelihood gap with autoregressive models of the same size, while achieving sample quality comparable to that of previous latent diffusion models.
  arXiv  Detail & Related papers  (2026-01-07T20:50:59Z)
• Breaking the Bottleneck with DiffuApriel: High-Throughput Diffusion LMs with Mamba Backbone [6.76700377196741]
  We introduce DiffuApriel, a masked diffusion language model built on a bidirectional Mamba backbone.<n>Our results demonstrate that bidirectional state-space architectures serve as strong denoisers in masked diffusion LMs.
  arXiv  Detail & Related papers  (2025-11-19T23:23:49Z)
• Improving Progressive Generation with Decomposable Flow Matching [50.63174319509629]
  Decomposable Flow Matching (DFM) is a simple and effective framework for the progressive generation of visual media.<n>On Imagenet-1k 512px, DFM achieves 35.2% improvements in FDD scores over the base architecture and 26.4% over the best-performing baseline.
  arXiv  Detail & Related papers  (2025-06-24T17:58:02Z)
• Discrete Diffusion in Large Language and Multimodal Models: A Survey [61.86669998363359]
  We provide a systematic survey of Discrete Diffusion Language Models (dLLMs) and Discrete Diffusion Multimodal Language Models (dMLLMs)<n>Unlike autoregressive (AR) models, dLLMs and dMLLMs adopt a multi-token, parallel decoding paradigm using full attention and a denoising-based generation strategy.<n>We trace the historical development of dLLMs and dMLLMs, formalize the underlying mathematical frameworks, list commonly-used modeling methods, and categorize representative models.
  arXiv  Detail & Related papers  (2025-06-16T17:59:08Z)
• Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective [54.77404771454794]
  We develop a flexible and robust world model for Multi-Agent Reinforcement Learning (MARL) using diffusion models.<n>Our method, Diffusion-Inspired Multi-Agent world model (DIMA), achieves state-of-the-art performance across multiple multi-agent control benchmarks.
  arXiv  Detail & Related papers  (2025-05-27T09:11:38Z)
• Dual Diffusion for Unified Image Generation and Understanding [32.7554623473768]
  We propose a large-scale and fully end-to-end diffusion model for multi-modal understanding and generation.<n>We leverage a cross-modal maximum likelihood estimation framework that simultaneously trains the conditional likelihoods of both images and text jointly.<n>Our model attained competitive performance compared to recent unified image understanding and generation models.
  arXiv  Detail & Related papers  (2024-12-31T05:49:00Z)
• Multimodal Latent Language Modeling with Next-Token Diffusion [111.93906046452125]
  Multimodal generative models require a unified approach to handle both discrete data (e.g., text and code) and continuous data (e.g., image, audio, video)<n>We propose Latent Language Modeling (LatentLM), which seamlessly integrates continuous and discrete data using causal Transformers.
  arXiv  Detail & Related papers  (2024-12-11T18:57:32Z)
• Energy-Based Diffusion Language Models for Text Generation [126.23425882687195]
  Energy-based Diffusion Language Model (EDLM) is an energy-based model operating at the full sequence level for each diffusion step.<n>Our framework offers a 1.3$times$ sampling speedup over existing diffusion models.
  arXiv  Detail & Related papers  (2024-10-28T17:25:56Z)
• Improving Fine-Grained Control via Aggregation of Multiple Diffusion Models [7.011853084733679]
  This paper introduces a novel training-free algorithm, independent of denoising network architectures, for fine-grained generation.<n>The algorithm integrates features from multiple diffusion models into a specified model to activate particular features and enable fine-grained control.<n> Experimental results demonstrate that AMDM significantly improves fine-grained control without training, validating its effectiveness.
  arXiv  Detail & Related papers  (2024-10-02T06:16:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.