Diffusion Language Model MMD Training Diagram
Overview
Three panels for training a diffusion language model with a maximum-mean-discrepancy objective. It is an editable PowerPoint diagram adapted from the original paper figure under CC BY 4.0.
What the diagram shows
Discrete training: corrupt → G_θ → token probabilities → G×B sampled sequences → MMD loss with a grouped REINFORCE gradient path. Continuous training: B noise inputs, K bootstrap passes through a frozen G_θ, then a trainable G_θ. A third panel computes a token-level RBF MMD² between reference and generated features from a frozen pretrained model. Forward and backward (dashed) arrows follow the legend. Every box, arrow, label and math term is a native shape you can recolor, resize and relabel for your own slides, posters or papers.
File format
The download contains .pptx (1 slide, 13.33 × 7.29 in; opens in PowerPoint, Keynote, WPS and LibreOffice Impress; Google Slides works too, but complex freeforms may shift slightly), .svg and .drawio. Includes preview.png (full resolution) and a 1200 px web preview.
Shapes and editing
The file has 173 native editable shapes: 94 auto-shapes, 43 text boxes, 17 freeforms, 19 lines/connectors (2 small groups). No embedded images. Everything is vector. You can edit sequence columns (𝐱₁…𝐱_L, 𝐳_{b,l}), G_θ blocks, the 4×3 probability grid, sampled-sequence cards, MMD boxes, forward/backward arrows with the stop-gradient ×, the legend and native flame / snowflake icons.
Embedded images and fonts
This file has no embedded images; everything is vector. Text uses Times New Roman, which is preinstalled on Windows and macOS (LibreOffice uses the metric-compatible Liberation Serif). A few calligraphic math letters (such as 𝒩, ℰ, 𝒮) use Cambria Math, which ships with Microsoft Office; other apps substitute a math font. Sub- and superscripts are real formatted text runs, not equation objects. The SVG keeps text as text; the draw.io file is a shape-by-shape conversion and may need small nudges. This is a faithful redraw for reuse, not the authors' original source file.
Source & license Adapted from Figure 1 of Representation-Space MMD for Diffusion Language Models by Ilya Drobyshevskiy, Ilia Sudakov, Maksim Semenov, Denis Kuznedelev, Maksim Ignatov, Pavel Temirchev, Nikita Balagansky, Viacheslav Meshchaninov, Nikita Gushchin, Dmitry Baranchuk (arXiv:2610.06648, https://arxiv.org/abs/2610.06648; code: https://github.com/yandex-research/dlm-mmd). Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Redrawn as editable shapes; changes made by 叫我Alonzo就好了/alonzolab. Not endorsed by the original authors. Attribution to the original paper is required when you publish or share the figure; the exact credit line is in LICENSE-ATTRIBUTION.txt inside each download.