Diffusion Models

LLaDA-Image combines a from-scratch 6B Diffusion Transformer with a frozen vision-language understanding module for generation, editing, and text rendering. Its distilled Turbo variant reduces inference to 2–4 sampling steps while the project releases weights, code, and training recipes.

EditBridge introduces a diffusion bridge for ultra-high-resolution image editing. Instead of independently super-resolving a low-resolution edit, it uses the original high-resolution image to guide detail recovery while reducing cross-image attention costs.

DiffusionGemma is an experimental open-weight language model that uses discrete diffusion to generate text in parallel blocks rather than token by token. Built by fine-tuning Gemma 4, it explores whether diffusion-style decoding can deliver a better speed-capability trade-off for LLMs.

This paper studies how structured language in prompts affects diffusion-based visual generation. Rather than treating prompt length as the key factor, it shows that measurable structure in text conditioning can correlate with lower converged diffusion loss.

Explorative Modeling reframes generative training by exploring multiple candidate matches between model outputs and data, then learning from the best one. The paper argues that exploration can scale alongside parameters and data.

Chimera proposes a hybrid visual diffusion architecture designed to make high-resolution image and long-video generation more compute-efficient. It combines linear-time state tracking, intermittent global interaction, local convolutions, sparse MoE capacity, and a scaling recipe called HeteroP.

A new paper introduces Parallel Decoding Distillation, a trajectory-based distillation method designed to speed up diffusion and flow-matching models without relying heavily on adversarial or VSD-style objectives.

Three-Body Scattering Modeling offers a new training signal for one-step generators by learning how generated samples should move toward real data, without an adversarial critic or a prescribed diffusion path.

Sony researchers introduce Spectral Alignment, a lightweight inference-time guidance method for correcting frequency-dependent mismatch in diffusion and flow-matching models. The method calibrates intermediate predictions with a precomputed spectral prior while adding only modest overhead.

This paper investigates whether diffusion language models internally track denoising progress. The authors find that a latent time-like signal can be decoded from residual streams even without explicit timestep conditioning.

A new paper argues that structural text template tokens in text-to-image Diffusion Transformers are not passive placeholders. They become implicit semantic registers that help preserve object identity during generation.

Mage-Flow argues that competitive image generation does not have to rely only on ever-larger parameter counts. Its 4B-scale stack combines an efficient latent tokenizer, a native-resolution diffusion transformer, and system-level CUDA optimizations.

DiFA is a training-free inference framework for diffusion models that treats intermediate predictions as correlated observations rather than disposable integration outputs. By building a forward-aligned temporal consensus, it aims to improve stability and preserve fine details during sampling.

DiffGI targets a long-standing weakness of many 3D generative models: representing thin-shell and non-manifold surfaces such as garments. By combining continuous 2D TSDF maps with differentiable Marching Squares, it links compact 2D latent representations directly to 3D surface optimization.

Token Time Continuous Diffusion introduces a continuous-space diffusion approach to language modeling, with different tokens progressing from noise to text at different rates. The method targets better high-speed generation, especially for conditional tasks.

A new survey argues that discrete diffusion should be understood through the way its state space is built. Tokenization, vocabulary topology, and domain alphabets shape everything from corruption to sampling.

A new arXiv paper introduces Cyclone, a latent diffusion framework for weather editing in autonomous driving scenes. The method aims to synthesize and remove adverse weather without requiring paired training data.

A new arXiv paper introduces HTFM, a flow-matching framework that models heavy-tailed sources as Gaussian mixtures conditioned on random clock paths. The goal is to make generative models better aligned with rare events, long-tail classes, and extreme-value data.

TCAM-Diff proposes a triplane-aware diffusion pipeline for high-resolution 3D medical image generation. Instead of modeling dense volumes directly, it learns compact triplane features and decodes them back into 3D volumes.

A new survey argues that discrete diffusion models should be understood through the construction of their state space, not merely through denoising objectives or samplers.