Papers5

#discrete diffusion

Balancing Understanding and Generation in Discrete Diffusion Models

This paper introduces XDLM, a single model that blends two popular diffusion styles (masked and uniform) so it both understands and generates text and images well.

#XDLM#discrete diffusion#stationary noise kernel

Diffusion In Diffusion: Reclaiming Global Coherence in Semi-Autoregressive Diffusion

Intermediate

Linrui Ma, Yufei Cui et al.Jan 20arXiv

The paper proposes Diffusion in Diffusion, a draft-then-revise method that brings back global coherence to fast, block-based diffusion language models.

#discrete diffusion#block diffusion#semi-autoregressive

Dream-VL & Dream-VLA: Open Vision-Language and Vision-Language-Action Models with Diffusion Language Model Backbone

Intermediate

Jiacheng Ye, Shansan Gong et al.Dec 27arXiv

Dream-VL and Dream-VLA use a diffusion language model backbone to understand images, talk about them, and plan actions better than many regular (autoregressive) models.

#diffusion language model#vision-language model#vision-language-action

Scaling Behavior of Discrete Diffusion Language Models

Intermediate

Dimitri von Rütte, Janis Fluri et al.Dec 11arXiv

This paper studies how a newer kind of language model, called a discrete diffusion language model (DLM), gets better as we give it more data, bigger models, and more compute.

#discrete diffusion#language models#scaling laws

Fast-Decoding Diffusion Language Models via Progress-Aware Confidence Schedules

Beginner

Amr Mohamed, Yang Zhang et al.Dec 2arXiv

Diffusion language models (dLLMs) can write all parts of an answer in parallel, but they usually take many tiny cleanup steps, which makes them slow.

#diffusion language models#early exit decoding#progress-aware threshold