# Zhijie Deng explains how to speed up diffusion language models without losing quality

> Original title: 上海交大邓志杰：扩散语言模型，怎样才能又快又好

- Source: Leiphone
- Published: 2026-11-05T05:44:00.000Z
- Bad Influence: https://badinfluence.news/items/z79p6mo8hiv13kdg54l0r5tpa
- Original: https://www.leiphone.com/category/private/T9k5sQixxLljsoH7.html

## Summary

Shanghai Jiao Tong University associate professor Zhijie Deng explained at CNCC 2026 how diffusion language models can generate faster while keeping output quality. He covered methods from LLaDA and Dream through cross-block parallel generation, post-training, and inference-system optimization. He measures efficiency by tokens committed per forward pass and speed in tokens per second, alongside answer and code-pass accuracy.