CDM migrates distribution matching distillation to continuous time via dynamic random-length schedules and active off-trajectory latent alignment, yielding competitive few-step image fidelity on SD3 and Longcat-Image.
hub Mixed citations
High- resolution image synthesis with latent diffusion models
Mixed citation behavior. Most common role is background (53%).
53
Pith papers citing it
Background
53% of classified citations
hub tools
citation-role summary
background 8
baseline 3
method 2
dataset 1
other 1
citation-polarity summary
representative citing papers
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
Cross-attention maps serve as a tractable surrogate for manifold proximity, enabling automatic synthesis of anchors that suppress normal-space drift in diffusion unlearning.
TILDE derives a minimum-deviation, energy-tilted target distribution for concept unlearning in diffusion models and realizes it via residual ∇-GFlowNet training.
Formalizes video world models as group actions on states and uses latent regularization with synthesized supervision to enforce consistency, introducing GAC and GAR metrics that improve structural correctness in SOTA models.
Loki replaces RGB conditioning stacks with identity-orthogonal parametric face encodings rasterized for diffusion, achieving efficient cross-ID portrait animation without cross-ID training data.
MDU minimizes forward KL divergence from prompt-conditional to prompt-masked unconditional predictions at masked positions to unlearn knowledge in MDLMs while trading off privacy and utility via temperature scaling.
MM-SOLD is a training-free particle sampler whose large-particle limit converges to a moment-matched Gibbs distribution obtained by exponentially tilting a score-smoothed target.
JEDI is the first online end-to-end latent diffusion world model that trains latents from denoising loss rather than reconstruction, achieving competitive Atari100k results with 43% less VRAM and over 3x faster sampling than pixel diffusion baselines.
SAEParate disentangles sparse representations in diffusion models via contrastive clustering and nonlinear encoding to enable more precise concept unlearning with reduced side effects.
DRoRAE adaptively fuses multi-layer features from vision encoders via energy-constrained routing to enrich visual tokens, cutting rFID from 0.57 to 0.29 and generation FID from 1.74 to 1.65 on ImageNet-256 while revealing a log-linear scaling law with fusion capacity.
PIQL integrates privileged information to accelerate convergence, lower loss, and improve generalization in tabular foundation models.
A training-free adaptive subspace projection method mitigates semantic collapsing in generative personalization by isolating and adjusting drift in a low-dimensional subspace using the stable pre-trained embedding as anchor.
LENS shapes low-frequency eigen noise with a lightweight network to enable efficient, high-quality sampling in distilled diffusion models.
ArenaPO infers Gaussian capability distributions from pairwise preferences and applies truncated-normal latent inference to derive fine-grained offline rewards for preference optimization of text-to-image diffusion models.
A feed-forward feature-Gaussian plus one-step geometry-aware pixel-flow simulator converts large image collections into 20K interactive scenes and 10M+ navigation samples that improve zero-shot Habitat and real-robot performance.
Point-cloud skeleton conditions and a Reset-and-Roll inference scheme enable stable frame-wise autoregressive driving video generation for closed-loop autonomous driving simulation.
Patch-PODiff-ViT defines a structured latent space via patchwise POD for efficient diffusion-based super-resolution and direct analytic uncertainty quantification across scientific and natural images.
GORMPO uses generative models for density-based regularization in model-based offline RL, outperforming baselines by 17% on a medical dataset while providing theoretical guarantees under mild assumptions.
A vanilla Diffusion Transformer trained via x-prediction on frozen DINOv2 features reaches FID 1.14 on ImageNet 256x256 with fewer parameters and faster sampling than prior DiT variants.
CPC-VAR adds Gradient-based Concept Neuron Selection for continual single-concept learning and a context-aware multi-branch composition strategy to reduce forgetting and entanglement in VAR-based personalized image generation.
SiLD is a score-matching framework that learns both manifold projection and intrinsic density from a single objective, with proven sample complexity depending only on intrinsic dimension.
SemaVoice adds SFM-guided alignment to refine continuous speech representations in autoregressive TTS, reporting 1.71% English WER on Seed-TTS and competitiveness with open-source SOTA.
BTECF encodes retinal vessels as Bézier trees to enable targeted, parameter-level counterfactual interventions on vessel geometry for causal analysis of vascular diseases.
citing papers explorer
-
Continuous-Time Distribution Matching for Few-Step Diffusion Distillation
CDM migrates distribution matching distillation to continuous time via dynamic random-length schedules and active off-trajectory latent alignment, yielding competitive few-step image fidelity on SD3 and Longcat-Image.
-
Flow-GRPO: Training Flow Matching Models via Online RL
Flow-GRPO is the first online RL method for flow matching models, raising GenEval accuracy from 63% to 95% and text-rendering accuracy from 59% to 92% with little reward hacking.
-
AutoAnchor: Stable Diffusion Unlearning Using Cross-Attention as a Manifold Surrogate
Cross-attention maps serve as a tractable surrogate for manifold proximity, enabling automatic synthesis of anchors that suppress normal-space drift in diffusion unlearning.
-
TILDE: TILt-based Distributional Erasure for Concept Unlearning
TILDE derives a minimum-deviation, energy-tilted target distribution for concept unlearning in diffusion models and realizes it via residual ∇-GFlowNet training.
-
World Models as Group Actions
Formalizes video world models as group actions on states and uses latent regularization with synthesized supervision to enforce consistency, introducing GAC and GAR metrics that improve structural correctness in SOTA models.
-
Loki: Representation over Architecture for Diffusion-Based Portrait Animation
Loki replaces RGB conditioning stacks with identity-orthogonal parametric face encodings rasterized for diffusion, achieving efficient cross-ID portrait animation without cross-ID training data.
-
Machine Unlearning for Masked Diffusion Language Models
MDU minimizes forward KL divergence from prompt-conditional to prompt-masked unconditional predictions at masked positions to unlearn knowledge in MDLMs while trading off privacy and utility via temperature scaling.
-
Training-Free Generative Sampling via Moment-Matched Score Smoothing
MM-SOLD is a training-free particle sampler whose large-particle limit converges to a moment-matched Gibbs distribution obtained by exponentially tilting a score-smoothed target.
-
JEDI: Joint Embedding Diffusion World Model for Online Model-Based Reinforcement Learning
JEDI is the first online end-to-end latent diffusion world model that trains latents from denoising loss rather than reconstruction, achieving competitive Atari100k results with 43% less VRAM and over 3x faster sampling than pixel diffusion baselines.
-
Disentangled Sparse Representations for Concept-Separated Diffusion Unlearning
SAEParate disentangles sparse representations in diffusion models via contrastive clustering and nonlinear encoding to enable more precise concept unlearning with reduced side effects.
-
Beyond the Last Layer: Multi-Layer Representation Fusion for Visual Tokenization
DRoRAE adaptively fuses multi-layer features from vision encoders via energy-constrained routing to enrich visual tokens, cutting rFID from 0.57 to 0.29 and generation FID from 1.74 to 1.65 on ImageNet-256 while revealing a log-linear scaling law with fusion capacity.
-
Toward Privileged Foundation Models:LUPI for Accelerated and Improved Learning
PIQL integrates privileged information to accelerate convergence, lower loss, and improve generalization in tabular foundation models.
-
Adaptive Subspace Projection for Generative Personalization
A training-free adaptive subspace projection method mitigates semantic collapsing in generative personalization by isolating and adjusting drift in a low-dimensional subspace using the stable pre-trained embedding as anchor.
-
LENS: Low-Frequency Eigen Noise Shaping for Efficient Diffusion Sampling
LENS shapes low-frequency eigen noise with a lightweight network to enable efficient, high-quality sampling in distilled diffusion models.
-
Arena as Offline Reward: Efficient Fine-Grained Preference Optimization for Diffusion Models
ArenaPO infers Gaussian capability distributions from pairwise preferences and applies truncated-normal latent inference to derive fine-grained offline rewards for preference optimization of text-to-image diffusion models.
-
Image2Sim: Scaling Embodied Navigation via Generative Neural Simulator
A feed-forward feature-Gaussian plus one-step geometry-aware pixel-flow simulator converts large image collections into 20K interactive scenes and 10M+ navigation samples that improve zero-shot Habitat and real-robot performance.
-
Point as Skeleton: Accumulated Point Cloud Enhanced Autoregressive Generation for Closed-Loop Autonomous Driving Simulation
Point-cloud skeleton conditions and a Reset-and-Roll inference scheme enable stable frame-wise autoregressive driving video generation for closed-loop autonomous driving simulation.
-
Patch-PODiff-ViT: Structured Latent Diffusion with Patchwise POD for Super-Resolution and Uncertainty Quantification
Patch-PODiff-ViT defines a structured latent space via patchwise POD for efficient diffusion-based super-resolution and direct analytic uncertainty quantification across scientific and natural images.
-
Generative OOD-regularized Model-based Policy Optimization
GORMPO uses generative models for density-based regularization in model-based offline RL, outperforming baselines by 17% on a medical dataset while providing theoretical guarantees under mild assumptions.
-
RiT: Vanilla Diffusion Transformers Suffice in Representation Space
A vanilla Diffusion Transformer trained via x-prediction on frozen DINOv2 features reaches FID 1.14 on ImageNet 256x256 with fewer parameters and faster sampling than prior DiT variants.
-
CPC-VAR:Continual Personalized and Compositional Generation in Visual Autoregressive Models
CPC-VAR adds Gradient-based Concept Neuron Selection for continual single-concept learning and a context-aware multi-branch composition strategy to reduce forgetting and entanglement in VAR-based personalized image generation.
-
Provably Learning Diffusion Models under the Manifold Hypothesis: Collapse and Refine
SiLD is a score-matching framework that learns both manifold projection and intrinsic density from a single objective, with proven sample complexity depending only on intrinsic dimension.
-
SemaVoice: Semantic-Aware Continuous Autoregressive Speech Synthesis
SemaVoice adds SFM-guided alignment to refine continuous speech representations in autoregressive TTS, reporting 1.71% English WER on Seed-TTS and competitiveness with open-source SOTA.
-
A General B\'ezier Tree Encoding Counterfactual Framework for Retinal-Vessel-Mediated Disease Analysis
BTECF encodes retinal vessels as Bézier trees to enable targeted, parameter-level counterfactual interventions on vessel geometry for causal analysis of vascular diseases.
-
OmniHumanoid: Streaming Cross-Embodiment Video Generation with Paired-Free Adaptation
OmniHumanoid factorizes transferable motion learning from embodiment-specific adaptation to enable scalable cross-embodiment video generation without paired data for new humanoids.
-
$h$-control: Training-Free Camera Control via Block-Conditional Gibbs Refinement
h-control augments hard-replacement guidance with block-conditional pseudo-Gibbs refinement on unobserved latent sites and adaptive 3D patch freezing to achieve superior FVD on RealEstate10K and DAVIS.
-
SocialDirector: Training-Free Social Interaction Control for Multi-Person Video Generation
SocialDirector uses spatiotemporal actor masking and directional reweighting on cross-attention maps to reduce actor-action mismatches and improve target-directed interactions in generated multi-person videos.
-
Toward Better Geometric Representations for Molecule Generative Models
LENSEs improves representation-conditioned molecule generation by jointly training a multi-level representation head, perceptual loss, and REPA alignment on pretrained encoders, yielding 97.28% validity and 98.51% stability on GEOM-DRUG.
-
Flow-Direct: Feedback-Efficient and Reusable Guidance for Flow Models via Non-Parametric Guidance Field
Flow-Direct constructs a reusable non-parametric guidance field from the log-density ratio of base and target distributions using all accumulated reward samples for feedback-efficient guidance in flow models.
-
SwiftI2V: Efficient High-Resolution Image-to-Video Generation via Conditional Segment-wise Generation
SwiftI2V achieves comparable 2K I2V quality to end-to-end models on VBench-I2V while cutting GPU time by 202x through low-resolution motion planning followed by strongly image-conditioned segment-wise high-resolution synthesis.
-
Taming Outlier Tokens in Diffusion Transformers
Outlier tokens in DiTs are addressed with Dual-Stage Registers, which reduce artifacts and improve image generation on ImageNet and text-to-image tasks.
-
Structured 3D Latents Are Surprisingly Powerful: Unleashing Generalizable Style with 2D Diffusion
DiLAST optimizes 3D latents via guidance from a 2D diffusion model to enable generalizable style transfer for OOD styles in 3D asset generation.
-
Post-Hoc Guidance for Consistency Models by Joint Flow Distribution Learning
JFDL allows pre-trained Consistency Models to perform guided image generation post-hoc by aligning flow distributions, reducing FID scores on CIFAR-10 and ImageNet without needing a teacher model.
-
Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
A 6B single-stream diffusion transformer trained with heavily curated data reaches top open-source image-generation quality in 314K H800 GPU hours, releasing Turbo and Edit variants.
-
HunyuanImage 3.0 Technical Report
HunyuanImage 3.0 is an open 80B-parameter multimodal autoregressive image generator that reportedly matches leading closed models on in-house benchmarks.
-
Slot-MLLM: Object-Centric Visual Tokenization for Multimodal LLM
Slot-MLLM introduces a slot-attention-based object-centric visual tokenizer with Q-Former encoder, diffusion decoder, and residual vector quantization for improved local visual comprehension and generation in multimodal LLMs.
-
Adversarial Error Correction for Visual Autoregressive Generation
AID-VAR attaches an adversarial discriminator and lightweight guidance injector to frozen VAR backbones to diagnose and correct fidelity gaps across scales, reporting 16% FID gains with 3% added parameters.
-
GenEvolve: Self-Evolving Image Generation Agents via Tool-Orchestrated Visual Experience Distillation
GenEvolve introduces a self-evolving agent framework for image generation using tool-orchestrated trajectories and Visual Experience Distillation to achieve claimed SOTA results on benchmarks.
-
Omni-Customizer: End-to-End MultiModal Customization for Joint Audio-Video Generation
Omni-Customizer proposes an end-to-end framework using Omni-Context Fusion, Masked TTS Cross-Attention, Semantic-Anchored Multimodal RoPE, and specialized training curricula to achieve precise multimodal identity binding in joint audio-video generation.
-
Registers Matter for Pixel-Space Diffusion Transformers
Register tokens improve pixel-space Diffusion Transformers by cleaning high-noise feature maps, and Register Guidance amplifies that effect.
-
HDRFace: Rethinking Face Restoration with High-Dimensional Representation
HDRFace injects high-dimensional facial features from low-quality and intermediate images into diffusion models via SDFM fusion, reporting gains on SD V2.1-base and Qwen-Image.
-
Composable Crystals: Controllable Materials Discovery via Concept Learning
VQ-VAE concept learning enables controllable recombination of crystal motifs to generate structures with reported gains in validity-stability-uniqueness-novelty metrics on MP-20 and Alex-MP-20.
-
Micro-Defects Expose Macro-Fakes: Detecting AI-Generated Images via Local Distributional Shifts
MDMF detects AI-generated images by learning patch-level forensic signatures and quantifying their distributional discrepancies with MMD, yielding larger separation than global methods when micro-defects are present.
-
PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation
PG-LRF generates signal-faithful and physiologically plausible ECGs from PPG inputs by structuring a latent space with an electro-hemodynamic simulator and enforcing consistency in a rectified flow model.
-
Diverse Sampling in Diffusion Models with Marginal Preserving Particle Guidance
EDDY adds diversity to diffusion-model samples by using kernel-based anti-symmetric pairwise drifts that preserve marginal distributions via Fokker-Planck symmetries, with practical approximations for expensive cases.
-
Steering Visual Generation in Unified Multimodal Models with Understanding Supervision
Using understanding tasks as direct supervision during post-training improves image generation and editing in unified multimodal models.
-
A Systematic Post-Train Framework for Video Generation
A post-training pipeline for video generation models combines SFT, RLHF with novel GRPO, prompt enhancement, and inference optimization to improve visual quality, temporal coherence, and instruction following.
-
The Amazing Stability of Flow Matching
Flow matching generative models preserve sample quality, diversity, and latent representations despite pruning 50% of the CelebA-HQ dataset or altering architecture and training configurations.
-
VRAG: Learning World Models for Interactive Video Generation
VRAG improves long-horizon interactive video generation by conditioning autoregressive diffusion on retrieved historical frames and explicit global state, outperforming long-context baselines on the tested Minecraft and RealEstate10K benchmarks.
-
Enhancing Text-to-Image Diffusion Transformer via Split-Text Conditioning
DiT-ST converts complete-text captions into split-text primitives via LLMs and injects them hierarchically across denoising stages to reduce semantic confusion in DiT-based text-to-image generation.