A fine-tuned video diffusion model translates monocular video into a synthetic proxy video of a moving cube, enabling 6-DoF pose tracking via classical solvers without 3D models, depth, or masks.
super hub Mixed citations
Classifier-Free Diffusion Guidance
Mixed citation behavior. Most common role is method (49%).
604
Pith papers citing it
19
external citations · Pith
Method
49% of classified citations
abstract
Classifier guidance is a recently introduced method to trade off mode coverage and sample fidelity in conditional diffusion models post training, in the same spirit as low temperature sampling or truncation in other types of generative models. Classifier guidance combines the score estimate of a diffusion model with the gradient of an image classifier and thereby requires training an image classifier separate from the diffusion model. It also raises the question of whether guidance can be performed without a classifier. We show that guidance can be indeed performed by a pure generative model without such a classifier: in what we call classifier-free guidance, we jointly train a conditional and an unconditional diffusion model, and we combine the resulting conditional and unconditional score estimates to attain a trade-off between sample quality and diversity similar to that obtained using classifier guidance.
hub tools
citation-role summary
method 53
background 49
baseline 5
extension 1
other 1
citation-polarity summary
claims ledger
- abstract Classifier guidance is a recently introduced method to trade off mode coverage and sample fidelity in conditional diffusion models post training, in the same spirit as low temperature sampling or truncation in other types of generative models. Classifier guidance combines the score estimate of a diffusion model with the gradient of an image classifier and thereby requires training an image classifier separate from the diffusion model. It also raises the question of whether guidance can be performed without a classifier. We show that guidance can be indeed performed by a pure generative model w
authors
co-cited works
representative citing papers
WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
Joint KL yields horizon-free approximation but an information-theoretic lower bound of order Omega(H) for estimation error in autoregressive learning, with matching computationally efficient upper bounds.
Inference-time refinement of pre-trained tabular diffusion models via Bidirectional Chamfer Refinement achieves median 8.6% better downstream performance than real data across 15 benchmarks while preserving fidelity and privacy.
ASTRA reframes transition-state search as guided diffusion inference that samples the isodensity surface between metastable basins and converges to first-order saddles via score differences and physical forces.
LLaDA is a scalable diffusion-based language model that matches autoregressive LLMs like LLaMA3 8B on tasks and surpasses GPT-4o on reversal poem completion.
SCOREBED isolates EIG double intractability in a policy-independent score-matching stage, then trains design policies with a singly intractable gradient estimator, enabling cheap multi-policy selection.
Replacing CFG's w(r-1) coefficient with r^(1+w)-r removes a sigma_min-divergent residual blow-up on a Gaussian calibration model and stabilizes high-guidance diffusion sampling at zero extra NFE.
D2PO learns better low-NFE diffusion timestep schedules and CFG weights via DPO on a score-based energy with a dynamic denser-schedule preference target.
Proposes diffeomorphic optimization for manifold-constrained problems in generative models via flow maps, with Lie-group extensions for protein design showing metric improvements.
MemLearner introduces a learning-based adaptive context query method using query tokens in video world models to improve long-term scene consistency over rule-based retrieval.
FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates like 6.25 Hz.
ReMatch corrects train-test residual distribution mismatch in probabilistic downscaling via optimal transport in low-dimensional PCA space, reducing under-dispersion and improving SSR and CRPS on HRRR-ERA5 wind data.
A temperature-conditioned diffusion model trained on small XY lattices produces accurate larger-lattice samples and cuts MCMC thermalization time by roughly 10x.
NeuWorld uses a transformer VAE to learn compact Neural Implicit Scenes from sparse posed frames and a diffusion transformer to evolve them conditioned on camera trajectories for consistent interactive exploration.
OvESyn is the first text-conditioned 3D CT synthesis framework for abdomino-pelvic oncologic imaging, constructing evidence-based text from metadata to adapt a latent diffusion model across the domain gap from chest CT pretraining.
DeVAR is the first application of visual autoregressive modeling to low-dose CT denoising, using next-scale token prediction, a residual refiner, and hybrid discrete-continuous decoding to outperform prior methods on two public datasets.
OrthoTryOn uses Orthogonal Subspace Projection on shared LoRA and Fisher-guided Negative Guidance to enable conflict-free unified fashion generation, outperforming task-specific models on benchmarks.
MLFMs combine masking with continuous flows to scale flow-based language models to reasoning and instruction-following tasks on GSM8K and MT-Bench.
SharpMoE is a plug-and-play post-training method that uses clean latent features and a trajectory routing loss to enable accurate saliency-based routing in diffusion MoE models for improved visual generation.
Introduces the MGI task for distinguishing training members from generated samples and proposes the DCB method combining autoencoder and latent signals to address failures in prior membership inference and attribution approaches.
A technique for controllable diversity in text-to-image generation by inducing structured semantic variations at the prompt level via VLM and agentic workflow.
IMAGIN-4D introduces spatio-temporal decomposition of image conditioning in a diffusion-based generator to produce controllable human-object interactions from text, waypoints, and a reference image.
DiT-Reward converts pretrained DiT models into reward predictors that outperform HPSv3 on four benchmarks while providing 1.65x inference speedup.
citing papers explorer
-
ProxyPose: 6-DoF Pose Tracking via Video-to-Video Translation
A fine-tuned video diffusion model translates monocular video into a synthetic proxy video of a moving cube, enabling 6-DoF pose tracking via classical solvers without 3D models, depth, or masks.
-
WavTTS: Towards High-Quality Zero-Shot TTS via Direct Raw Waveform Modeling
WavTTS is the first raw-waveform diffusion TTS model using DiT flow matching and multi-scale mel supervision that approaches SOTA latent zero-shot performance while beating prior end-to-end models.
-
Autoregressive Learning in Joint KL: Sharp Oracle Bounds and Lower Bounds
Joint KL yields horizon-free approximation but an information-theoretic lower bound of order Omega(H) for estimation error in autoregressive learning, with matching computationally efficient upper bounds.
-
Inference-Time Refinement Closes the Synthetic-Real Gap in Tabular Diffusion
Inference-time refinement of pre-trained tabular diffusion models via Bidirectional Chamfer Refinement achieves median 8.6% better downstream performance than real data across 15 benchmarks while preserving fidelity and privacy.
-
A Priori Sampling of Transition States with Guided Diffusion
ASTRA reframes transition-state search as guided diffusion inference that samples the isodensity surface between metastable basins and converges to first-order saddles via score differences and physical forces.
-
Large Language Diffusion Models
LLaDA is a scalable diffusion-based language model that matches autoregressive LLMs like LLaMA3 8B on tasks and surpasses GPT-4o on reversal poem completion.
-
Bayesian Experimental Design via Score Matching
SCOREBED isolates EIG double intractability in a policy-independent score-matching stage, then trains design policies with a singly intractable gradient estimator, enabling cheap multi-policy selection.
-
Guidance Breaks the Fitted Operator: A Terminal-Fitted Repair for Classifier-Free Guidance
Replacing CFG's w(r-1) coefficient with r^(1+w)-r removes a sigma_min-divergent residual blow-up on a Gaussian calibration model and stabilizes high-guidance diffusion sampling at zero extra NFE.
-
D2PO: Optimizing Diffusion Samplers via Dynamic Preference
D2PO learns better low-NFE diffusion timestep schedules and CFG weights via DPO on a score-based energy with a dynamic denser-schedule preference target.
-
Diffeomorphic Optimization
Proposes diffeomorphic optimization for manifold-constrained problems in generative models via flow maps, with Lie-group extensions for protein design showing metric improvements.
-
MemLearner: Learning to Query Context memory for Video World Models
MemLearner introduces a learning-based adaptive context query method using query tokens in video world models to improve long-term scene consistency over rule-based retrieval.
-
FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model
FlexiSLM is the first spoken language model supporting dynamic and controllable frame rates on speech input and output, outperforming fixed-rate 7B models at high quality and enabling faster inference at lower rates like 6.25 Hz.
-
Mind the Residual Gap: Probabilistic Downscaling under Real-World Bias
ReMatch corrects train-test residual distribution mismatch in probabilistic downscaling via optimal transport in low-dimensional PCA space, reducing under-dispersion and improving SSR and CRPS on HRRR-ERA5 wind data.
-
Diffusion-warm sampling of the XY model enables fast thermalization at scale
A temperature-conditioned diffusion model trained on small XY lattices produces accurate larger-lattice samples and cuts MCMC thermalization time by roughly 10x.
-
Walking in the Implicit: Interactive World Exploration via Neural Scene Representation
NeuWorld uses a transformer VAE to learn compact Neural Implicit Scenes from sparse posed frames and a diffusion transformer to evolve them conditioned on camera trajectories for consistent interactive exploration.
-
Evidence-Based Text-Conditioned 3D CT Synthesis for Ovarian Cancer
OvESyn is the first text-conditioned 3D CT synthesis framework for abdomino-pelvic oncologic imaging, constructing evidence-based text from metadata to adapt a latent diffusion model across the domain gap from chest CT pretraining.
-
DeVAR: Low-Dose CT Denoising via Visual Autoregressive Modeling
DeVAR is the first application of visual autoregressive modeling to low-dose CT denoising, using next-scale token prediction, a residual refiner, and hybrid discrete-continuous decoding to outperform prior methods on two public datasets.
-
OrthoTryOn: Geometric Orthogonalization for Conflict-Free Unified Fashion Generation
OrthoTryOn uses Orthogonal Subspace Projection on shared LoRA and Fisher-guided Negative Guidance to enable conflict-free unified fashion generation, outperforming task-specific models on benchmarks.
-
Masked Language Flow Models
MLFMs combine masking with continuous flows to scale flow-based language models to reasoning and instruction-following tasks on GSM8K and MT-Bench.
-
Focusing on What Matters: Saliency-Harnessing Accurate Routing for Diffusion MoE
SharpMoE is a plug-and-play post-training method that uses clean latent features and a trajectory routing loss to enable accurate saliency-based routing in diffusion MoE models for improved visual generation.
-
MGI: Member vs Generated Inference
Introduces the MGI task for distinguishing training members from generated samples and proposes the DCB method combining autoencoder and latent signals to address failures in prior membership inference and attribution approaches.
-
Semantic Browsing: Controllable Diversity for Image Generation
A technique for controllable diversity in text-to-image generation by inducing structured semantic variations at the prompt level via VLM and agentic workflow.
-
IMAGIN-4D: Image-Guided Controllable Interaction Generation
IMAGIN-4D introduces spatio-temporal decomposition of image conditioning in a diffusion-based generator to produce controllable human-object interactions from text, waypoints, and a reference image.
-
DiT-Reward: Generative Representations for Text-to-Image Reward Modeling
DiT-Reward converts pretrained DiT models into reward predictors that outperform HPSv3 on four benchmarks while providing 1.65x inference speedup.
-
C^2GR: Coupled Comprehensive Generative Replay for a Continually Learnable Universal Segmentation Model
C^2GR uses Bayesian Joint Diffusion for coupled image-mask synthesis and Relation-aware Unified Prompt Synchronization to reduce forgetting in continual universal segmentation, showing a 2.44% performance drop versus joint training on 20 tasks.
-
Safe Few-Step Generation via Velocity Editing
VESFlow edits the learned velocity field of flow matching models via a safe-conditional posterior to produce safe images in 4 sampling steps, with an optional risk filter and VESFlow+ variant that also repels from unsafe directions.
-
Bagpiper-TTS: Natural Language Guided Universal Speech Synthesis
Bagpiper-TTS uses natural language prompts and intent reasoning to derive rich captions that guide a single model for universal speech synthesis across classical TTS, multi-talker, singing, and role-play tasks.
-
One-Prompt Censorship Evasion via Generative Diffusion Models
FlowPaint reframes censorship evasion as semantic image-to-image editing with instruction-tuned diffusion models, claiming single-prompt robustness against both rule-based and ML-based censors.
-
Variance-Tilted Diffusion Models for Diverse Sampling
Derives an interacting-particle diffusion sampler via Doob h-transform that targets batches with high empirical variance under a linear feature map.
-
Energy-based Compositional Diffusion Planning
ECD reformulates compositional diffusion planning as energy minimization over local bridge potentials, adding a boundary reaction term and a Markov score approximation that runs in linear time.
-
Context-Aware Autoregressive Diffusion for Gloss-Wise Sign Language Production
GARD is a context-aware autoregressive diffusion model for gloss-wise sign language production using inter-gloss transition guidance and global motion harmonizer, claiming superior linguistic accuracy and motion similarity on Phoenix-T and CSL-Daily datasets.
-
FlowBender: Feedback-Aware Training for Self-Correcting Conditional Flows
FlowBender introduces closed-loop training that lets conditional flow models learn correction policies from their own task-specific alignment errors, outperforming supervised and guidance baselines on fidelity and plausibility.
-
Scaling Generative Foundation Models for Chest Radiography with Rectified Flow Transformers
A 1.3B-parameter rectified flow transformer is the first generative foundation model for chest radiograph synthesis at billion-parameter scale, producing images indistinguishable from real ones to experts.
-
Image Prompt Reconstruction Attacks on Distributed MLLM Inference Frameworks
First study of image prompt reconstruction attacks on distributed MLLM inference, proposing MPAA for pixel-level and IEDA for semantic reconstruction with 100% embedding extraction accuracy on four model families.
-
Constrained Diffusion Models with Primal-Dual Inference
Develops primal-dual inference (PDI) that jointly infers optimal primal distributions and dual multipliers during diffusion sampling using a dual-conditioned score network.
-
Towards More General Control of Diffusion Models Using Jeffrey Guidance
Jeffrey guidance applies Jeffrey's rule of conditioning to diffusion models to target prescribed marginal distributions while preserving conditional structure, demonstrated via embedding matching and fairness enforcement.
-
M*: A Modular, Extensible, Serving System for Multimodal Models
M* introduces the Walk Graph abstraction to serve arbitrary compositions of multimodal model components and reports latency and throughput gains over vLLM-Omni and other baselines on text-to-image, text-to-speech, and robotic planning workloads.
-
Ambient Diffusion Policy: Imitation Learning from Suboptimal Data in Robotics
Ambient Diffusion Policy enables better imitation learning from suboptimal robot data by leveraging spectral properties to restrict data usage to specific diffusion times.
-
TopoCap: Learning Topology-Agnostic Motion Priors for Monocular Video-to-Animation
A two-stage generative model (Graph CVAE + flow matching) learns topology-agnostic motion codes from a new 5k-topology dataset and retargets video motion to arbitrary unseen skeletons.
-
VFUSE: Virulent Feature Understanding with Sparse autoEncoders
VFUSE applies sparse autoencoders to diffusion-transformer activations in RoseTTAFold3 and RFDiffusion3 to find monosemantic features that detect hazardous protein designs with AUROC up to 0.84.
-
Learning Where to Simulate: Generative Active Sampling for Online PDE Surrogate Training
OGAS uses a parallel diffusion model to bias PDE configuration sampling toward high surrogate difficulty, reducing 99th-percentile errors and error variance versus uniform sampling across tested 2D PDEs.
-
Continuous Language Diffusion as a Decoder-Interface Problem
Continuous language diffusion works by entering high-margin decoder basins where frozen T5 embeddings recover 93-96% of native decisions and linear readouts reach 97.9% agreement, implying models should be evaluated as representation-decoder systems.
-
Improving Bayesian Optimization via Training-Aware Conditional Diffusion Models
Conditional diffusion models trained with BO-aware strategies approximate the optimum distribution, enabling a Diffusion-based Mode Seeking acquisition function with a sub-optimality guarantee that outperforms baselines in experiments.
-
Complexity-Balanced Diffusion Splitting
CBS partitions the diffusion timeline into segments of equal approximation burden via Dirichlet energy and trajectory acceleration monitors estimated by an auxiliary model, yielding higher synthesis quality at fixed per-step cost across SiT, JiT and UNet backbones.
-
Physics in 2-Steps: Locking Motion Priors Before Visual Refinement Erases Them
PhaseLock extracts motion priors from 2-step inference and enforces them via Latent Delta Guidance to raise physical consistency scores by 6.2 points on average in image-to-video diffusion models.
-
Reflection Separation from a Single Image via Joint Latent Diffusion
A joint latent diffusion model with cross-layer self-attention and disjoint sampling separates reflection and transmission layers from single images more effectively than prior methods on real-world benchmarks.
-
Optimal Transport Flow Matching by Design
By designing the prior as the low-frequency projection of data images, flow matching achieves OT-optimal identity couplings without explicit OT computation, reducing trajectory curvature over 2x and improving few-step quality.
-
Diffusing in the Right Space: A Systematic Study of Latent Diffusability
A large-scale empirical study across tokenizers and diffusion backbones identifies Velocity Irreducible Variance (VIV) as one of the most stable predictors of latent diffusion generation quality.
-
JAVEDIT: Joint Audio-Visual Instruction-Guided Video Editing with Agentic Data Curation
JAVEdit-100k is the first large-scale dataset for instruction-guided joint audio-visual video editing, accompanied by JAVEditBench and the JAVEdit model that outperforms baselines on five of six metrics.
-
Drifting Preference Optimization for One-Step Generative Models
DrPO enables online preference optimization for deterministic one-step generators via non-parametric dipole updates from ranked samples plus base-model drift, without reward backpropagation.