Diffusion Models
PulseAugur coverage of Diffusion Models — every cluster mentioning Diffusion Models across labs, papers, and developer communities, ranked by signal.
7 天有情绪数据
-
New StAD method speeds up generative model likelihood calculations
Researchers have developed a new method called StAD to improve the speed and accuracy of likelihood calculations in diffusion and flow-based generative models. This technique bypasses the need to compute the Jacobian of…
-
New frameworks boost precipitation nowcasting with Mamba and diffusion models
Researchers have developed two new frameworks, MambaRain and VMU-Diff, to improve precipitation nowcasting accuracy for the crucial 0-3 hour window. MambaRain integrates Mamba's efficient long-range temporal modeling wi…
-
AnyBand-Diff framework enhances remote sensing image generation with spectral priors
Researchers have developed AnyBand-Diff, a new framework for generating and repairing remote sensing images. This model addresses limitations in existing diffusion models by incorporating spectral priors to ensure physi…
-
New AID method improves image inpainting with diffusion models
Researchers have developed a new method called Amortized Inpainting with Diffusion (AID) for image inpainting using pretrained diffusion models. AID trains a small, reusable guidance module offline, which can then be ap…
-
FlowSR achieves single-step image super-resolution with diffusion models
Researchers have developed FlowSR, a new method for image super-resolution that significantly speeds up the process using diffusion models. This approach reformulates super-resolution as a rectified flow from low-resolu…
-
He Kai Ming's team advances flow matching for faster image generation
He Kai Ming's team has published several papers challenging the dominance of diffusion models in image generation, proposing flow matching as a more efficient alternative. Their work introduces methods like JiT, which d…
-
Masked Generative Transformers offer faster, more precise image editing
Researchers have introduced EditMGT, a novel image editing framework utilizing Masked Generative Transformers (MGTs) as an alternative to dominant diffusion models. This MGT-based approach offers localized token predict…
-
Diffusion models and NeRF combine for probabilistic 3D scene reconstruction
Researchers have developed a novel method for 3D scene reconstruction by integrating diffusion models with Neural Radiance Fields (NeRF). This approach treats 3D reconstruction as a probabilistic problem, using a stocha…
-
New diffusion model enhances MRI reconstruction and coil sensitivity estimation
Researchers have developed a new method for reconstructing magnetic resonance images (MRIs) using diffusion models, which are known for generating high-quality images. This approach addresses limitations of existing tec…
-
Generative models learn rules across two distinct training timescales
Researchers have identified two distinct timescales in generative model training: the point at which generations become rule-valid ($\tau_{\mathrm{rule}}$) and the point at which models begin reproducing training sample…
-
New generative models leverage Wasserstein flows for faster, higher-quality outputs
Researchers are exploring new methods for generative modeling, focusing on Wasserstein gradient flows to improve efficiency and sample quality. One approach, W-Flow, achieves state-of-the-art one-step generation for ima…
-
AI research explores diffusion models, math agents, reasoning, and developer tools
A new research paper challenges existing understandings of diffusion models, suggesting a re-evaluation of their generalization properties and offering insights for future research directions in generative AI. Separatel…
-
Federated generative models analyzed for industrial predictive maintenance
A new research paper explores the use of generative models like VAEs, GANs, and Diffusion Models within federated learning frameworks for predictive maintenance in industrial settings. The study analyzes performance and…
-
New STMD method speeds diffusion model inference without teacher
Researchers have developed Stochastic Transition-Map Distillation (STMD), a novel framework designed to accelerate the inference process for diffusion models without requiring a pre-trained teacher model. This method di…
-
New diffusion model techniques accelerate video restoration and image sampling
Researchers have developed new methods to improve diffusion models for various inverse problems. One approach, AVIS, uses autoregressive diffusion models to accelerate video restoration, significantly reducing latency a…
-
Diffusion models offer new path for semantic communications in 6G
A new tutorial paper explores the integration of diffusion models, a type of generative AI, into semantic communication systems for 6G and beyond. It provides a comprehensive guide connecting diffusion techniques to com…
-
OpenAI tests ChatGPT ads in Japan, diffusion models improve code generation
OpenAI is testing advertisements within ChatGPT in Japan, targeting users of both free and 'Go' plans. This initiative aims to expand OpenAI's monetization strategies into the Japanese market. Separately, researchers ar…
-
TRACE framework enhances conformal prediction with diffusion and flow matching
Researchers have introduced TRACE, a novel framework for conformal prediction designed to handle multi-dimensional outputs. This method defines nonconformity by aligning transport dynamics within diffusion and flow matc…
-
New CDM method enhances diffusion model distillation for faster, higher-fidelity image generation
Researchers have introduced Continuous-Time Distribution Matching (CDM), a novel method for accelerating diffusion models. This approach moves beyond discrete-time distillation by employing a dynamic, continuous schedul…
-
New Eulerian Motion Guidance improves image animation with bidirectional consistency
Researchers have introduced Eulerian Motion Guidance, a novel method for animating static images using diffusion models. This approach shifts from traditional Lagrangian motion guidance to an Eulerian framework, utilizi…