Pseudo Numerical Methods for Diffusion Models on Manifolds

cs.LG 2025-04 conditional novelty 7.0 of 10

Diffusion model generalization is explained by the covariance of the noisy score-matching target, which adds state-dependent noise to sampling and fills gaps in the training distribution.

cs.CV 2026-08 conditional novelty 6.0 of 10

RippleNet detects AI-generated images by encoding multi-directional, multi-scale local pixel differences and using attention over those differences, achieving state-of-the-art cross-generator accuracy in the reported ...

cs.CV 2026-08 conditional novelty 6.0 of 10

Conditioning each patch's denoising on the full trajectories of earlier patches lets XYZFlow generate ImageNet images with FID 1.22 to 1.63 in only 2 to 5 steps per patch, at 7.2 to 8.5x teacher speedups.

cs.CV 2026-08 conditional novelty 6.0 of 10

LHSDet detects AI-generated images by fusing low-level texture features, SigLIP2 global features, and BLIP-2 captions in a visual question answering framework, reaching 96% average cross-model accuracy.

cs.CR 2026-07 accept novelty 6.0 of 10

One-step adversarially LoRA-tuned inversion exploits low-curvature reverse trajectories to beat 50-step DDIM on watermark robustness after ~20 minutes of fine-tuning.

cs.CV 2026-07 conditional novelty 6.0 of 10

Injecting spatial importance maps (e.g., CFG delta) into the OBS Hessian improves subject preservation in pruned diffusion models at high sparsity, but gains over the baseline are small and without error bars.

cs.CV 2026-07 conditional novelty 6.0 of 10

A conditional diffusion model maps Sentinel-2 optical imagery to 1 m sea-ice roughness residuals with ~9 cm RMSE on an unseen region, though pointwise correlation is weak (ZNCC≈0.11).

cs.LG 2026-04 unverdicted novelty 6.0 of 10

Ontology-aware contrastive pretraining plus note-to-vitals distillation improves MIMIC ICU risk and length-of-stay prediction using only vital signs at inference.

cs.CV 2025-08 conditional novelty 6.0 of 10

Grouping related text prompts into a tree and sharing early denoising steps with averaged embeddings saves 50 to 74 percent of diffusion compute on image-embedding-conditioned models while keeping VQA quality essentia...

cs.CV 2025-08 conditional novelty 6.0 of 10

A training-free extension of InteractDiffusion that uses LLM-mined relations, action feature offsets, and entity masks to improve entity and interaction control in generated images.

cs.CV 2025-07 conditional novelty 6.0 of 10

MPQ-DMv2 adds binary residual quantization, temporal relation distillation, and SVD-initialized LoRA to mixed-precision quantization, improving low-bit diffusion model generation quality.

cs.CV 2025-07 conditional novelty 6.0 of 10

A diffusion-based portrait style transfer method that uses semantic face alignment and an AdaIN-Wavelet latent blend to transfer style across photo, cartoon, sketch, and animation domains while preserving identity.

cs.CV 2025-06 conditional novelty 6.0 of 10

By training a semantic expert and a LoRA-based detail expert, DCM reaches nearly teacher-level VBench scores with 4-step video sampling on HunyuanVideo and CogVideoX.

cs.LG 2025-05 conditional novelty 6.0 of 10

The diffusion Fisher matrix of a Gaussian-perturbed distribution is expressed in the span of data outer products, enabling two faster approximation algorithms for trace and matrix-vector access.

cs.LG 2025-05 conditional novelty 6.0 of 10

A mix of signed and unsigned 4-bit floating-point formats, timestep-aware LoRA experts, and a denoising-weighted loss keeps diffusion-model image quality close to full precision.

cs.CV 2025-05 conditional novelty 6.0 of 10

Autoregressive image diffusion models can use far fewer denoising steps for later tokens without losing quality, yielding 1.4-2.5x speedup from step annealing and up to 10x when combined with fewer autoregressive steps.

cs.CV 2025-02 conditional novelty 6.0 of 10

MR Sampler derives semi-analytical solutions to the reverse-time SDE and PF-ODE of Mean Reverting Diffusion, enabling 5-10 step high-quality sampling instead of hundreds.

cs.CV 2025-02 conditional novelty 6.0 of 10

MaGNeTS schedules progressively larger nested transformer sub-models over decode iterations and caches key-value pairs of unmasked tokens, achieving 2.5-3.7x compute reduction with competitive FID/FVD.

cs.AR 2025-01 conditional novelty 6.0 of 10

Diffusion model inference can be accelerated by computing only the small differences between adjacent time steps, with reduced bit-width and zero-skipping, at moderate accuracy loss.

cs.CV 2025-01 conditional novelty 6.0 of 10

MObI jointly inpaints camera and lidar views of driving scenes, inserting objects from a single reference image at a user-specified 3D bounding box.

cs.CV 2025-01 conditional novelty 6.0 of 10

Token-based language-style vision models (VAR, LlamaGen) tolerate quantization better than diffusion models, and a custom TopKLD distillation loss pushes their low-bit scaling roughly one precision level higher.

cs.CV 2024-12 conditional novelty 6.0 of 10

DaTo combines dynamics-aware token pruning with feature caching and a searched per-step schedule, reaching 9x faster Stable Diffusion inference without FID loss.

cs.CV 2024-12 conditional novelty 6.0 of 10

A stochastic-start ODE sampler for diffusion bridge models avoids the singular start of the probability-flow ODE and beats prior samplers with fewer neural network evaluations.

cs.CV 2024-12 conditional novelty 6.0 of 10

An inversion-free, training-free diffusion editing method that anchors output latents to a pixel-manipulated copy of the image achieves consistent object repositioning, resizing, and pasting in 16 steps.

cs.LG 2024-12 conditional novelty 6.0 of 10

Learned diffusion score fields behave like Gaussian score fields at high noise, enabling an analytical 'teleportation' that skips early sampling steps without hurting FID.

cs.CV 2024-12 conditional novelty 6.0 of 10

UniPaint unifies video inpainting, outpainting, and frame interpolation in a single diffusion-based model with MoE attention and mixed-mask training.

cs.CV 2024-12 conditional novelty 6.0 of 10

BiDM achieves the first fully binarized (W1A1) diffusion model, with FID 22.74 on LSUN-Bedrooms, by adding timestep-aware binary structures and patch-based distillation.

cs.CV 2024-12 conditional novelty 6.0 of 10

A prompt-adaptive model selects the minimum denoising step count that reaches perceptual quality saturation, saving about 5 seconds per image with no detectable quality loss.

cs.CV 2024-12 conditional novelty 6.0 of 10

Head Relevance Vectors, built by counting which attention head best matches each of 34 concept words, reveal concept-specific cross-attention head patterns and enable targeted concept steering in Stable Diffusion.

cs.CV 2024-11 conditional novelty 6.0 of 10

CoDe speeds up Visual Auto-Regressive image generation by using a 2B model for early coarse scales and a 0.3B model for later fine scales, with 1.7x-2.9x speedup and only a small FID increase.

cs.CV 2024-11 conditional novelty 6.0 of 10

A token-merging method that anchors computation on high-CFG-importance tokens improves generation quality at fixed inference speedups.

cs.CV 2024-11 conditional novelty 6.0 of 10

A gradient-free diffusion sampler that selects a reference class from a pretrained classifier's predictions and adapts guidance strength improves class-conditional fidelity, but its Precision gains are measured with t...

cs.CV 2024-11 conditional novelty 6.0 of 10

A multi-stage text-to-image approach that generates individual objects as RGBA images and composes them scene-by-scene via noise blending, enabling fine-grained layout and attribute control.

cs.CV 2026-07 conditional novelty 5.5 of 10

Noise-perturbed condition injection plus contrastive trajectory refinement improves training-free conditional diffusion sampling across style transfer, super-resolution and deblurring.

cs.LG 2025-08 conditional novelty 5.0 of 10

VarDiU optimizes a variational upper bound on the diffusive KL divergence with an unbiased gradient estimator, improving one-step generation on a 2D 40-Gaussian toy benchmark compared with Diff-Instruct.

cs.CV 2025-08 unverdicted novelty 5.0 of 10

TimeMachine proposes a diffusion model with age-aware cross-attention and a latent age classifier, plus a 1M-image HFFA dataset, claiming SOTA age editing with identity preservation.

cs.AR 2025-07 conditional novelty 5.0 of 10

Phase-aware sampling cuts Stable Diffusion's compute by roughly 2.4x to 5.7x with only small CLIP-score changes, and the accompanying FPGA accelerator turns this into 2.7x to 6.0x energy savings over an Nvidia V100 GPU.

cs.CV 2025-06 conditional novelty 5.0 of 10

A cost-aware router selects per prompt the best among nine pre-trained text-to-image models, beating every single model on the quality-versus-cost frontier.

cs.LG 2025-06 conditional novelty 5.0 of 10

A weighted-particle sampler evolves the posterior through the diffusion model's reverse dynamics, with theoretical error bounds and improved image reconstructions.

cs.CV 2025-05 reject novelty 5.0 of 10

A training-free 'Levenberg-Marquardt-Langevin' diffusion sampler is claimed to improve image FID, but its update rule collapses to that of the baseline DPM-Solver for the parameter values used in the paper.

cs.CV 2025-05 conditional novelty 5.0 of 10

Training a diffusion model on reversed preference pairs and using it as the negative CFG predictor improves human-preference alignment over standard preference-tuned models.

cs.CV 2025-04 conditional novelty 5.0 of 10

GTF is a training-free, projection-based noise composition rule that enables text-driven addition, removal, and style transfer in diffusion models across image, video, and 3D generation.

cs.CV 2025-01 conditional novelty 5.0 of 10

Modeling quantization noise in compressed diffusion models as a time-step-dependent joint Gaussian, then correcting its mean and variance during sampling, improves FID over prior PTQ methods and can beat the full-prec...

cs.CV 2025-01 conditional novelty 5.0 of 10

CA-ToMe speeds up Stable Diffusion inference by adaptively merging similar tokens and caching merge pairs, reaching 6.09s average generation time with FID 34.05 on ImageNet.

cs.LG 2024-12 conditional novelty 5.0 of 10

LazyDiT learns small gates that decide when to reuse cached layer outputs, cutting diffusion transformer compute by up to half while matching or beating DDIM quality.

cs.SE 2025-12 reject novelty 4.0 of 10

A literature survey of scenario-generation methods for ADS testing that adds an unvalidated AII/RAS/OCS metric suite and ODD-difficulty schema, undermined by inconsistent calculations in the worked examples.

cs.LG 2025-06 conditional novelty 4.0 of 10

Joint training of autoencoders with diffusion models in latent space gives stochastic turbulence closure accuracy close to physical-space diffusion models at roughly 5-7x lower cost.

cs.CV 2025-06 conditional novelty 4.0 of 10

Dive3D shows that replacing KL divergence with score implicit matching in text-to-3D distillation, together with a reward term, produces more diverse and higher-fidelity 3D assets than SDS and ProlificDreamer baselines.

cs.CV 2025-06 conditional novelty 4.0 of 10

A training-free correction that blends each step's noise estimate with the initial noise estimate improves few-step diffusion sampling across DDIM, DPM-Solver, and DPM-Solver++.

cs.LG 2025-05 conditional novelty 4.0 of 10

A latent diffusion model with a context-aware autoencoder generates more diverse and synchronized listener facial reactions on REACT2024 than four comparison methods, at the cost of a slightly worse reaction-distance score.

eess.IV 2025-02 reject novelty 4.0 of 10

A new watermarking scheme for diffusion models trains an autoencoder to embed and recover image watermarks through the generation process, but the reported robustness is undermined by flawed evaluation and an unjustif...

cs.CV 2025-01 reject novelty 4.0 of 10

Combining ControlNet and GLIGEN, ObjectDiffusion conditions Stable Diffusion on bounding boxes and open-ended object names, reporting improved AP50, AR, and FID on COCO2017.

cs.CV 2024-12 conditional novelty 4.0 of 10

A serial-to-parallel distillation pipeline with per-timestep activation quantization makes W4A8 quantized Stable Diffusion models generate images much closer to the full-precision model than prior PTQ methods.

cs.CV 2024-11 conditional novelty 4.0 of 10

A k-step search over initial noises reduces quality drift in autoregressive chunk-by-chunk video generation, especially for small image-to-video models.

cs.CV 2025-08 conditional novelty 3.0 of 10

A DDPM-trained noise predictor can denoise across several time steps in one update because the multi-step posterior is Gaussian and uses the same network.

cs.CV 2025-05 conditional novelty 3.0 of 10

A structured review and benchmark of methods for quantizing diffusion models, with a taxonomy of post-training and quantization-aware approaches and an analysis of quantization artifacts.

cs.CV 2024-12