الملخص
Masked diffusion models (MDMs) generate sequences by progressively unmasking several tokens per denoising step, but their reverse process is typically factorized over positions, limiting sample quality in the few-step regime where diffusion's speed advantage over autoregressive decoding matters most. A recent line of work introduces a continuous Gaussian latent, trained as a variational autoencoder, to capture correlations across positions, but such approaches are prone to posterior collapse, where the latent is silently ignored. We propose Enhanced Mixture-of-Experts (E-MoE), which builds the reverse process as a mixture of factorized distributions over a discrete shared latent given by the expert-routing decisions of a Mixture-of-Experts (MoE) backbone, without increasing active parameters over the factorized baseline. Across synthetic multi-modal benchmarks, binarized MNIST, and LM1B, E-MoE improves few-step generation over factorized baselines.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Ivanov, A., Kolesov, A., Korotin, A., Oseledets, I., & Goncharov, M. (2026). E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models. https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models
MLA 9
Ivanov, Arseny, et al. "E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models." https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models.
شيكاغو (المؤلف–التاريخ)
Ivanov, Arseny, Alexander Kolesov, Alexander Korotin, Ivan Oseledets, and Mikhail Goncharov. 2026. "E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models." https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models.
هارفارد
Ivanov, A., Kolesov, A., Korotin, A., Oseledets, I. and Goncharov, M. (2026) 'E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models', Available at: https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models.
فانكوفر
Ivanov A, Kolesov A, Korotin A, Oseledets I, Goncharov M. E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models. https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models
IEEE
A. Ivanov, A. Kolesov, A. Korotin, I. Oseledets, and M. Goncharov, "E-MoE: Enhanced Mixture-of-Experts for Non-Factorized Diffusion Language Models," https://omanscience.com/ar/articles/e-moe-enhanced-mixture-of-experts-for-non-factorized-diffusion-language-models.