REVIEW 29 cited by
Distribution Matching Distillation Meets Reinforcement Learning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Distribution Matching Distillation Meets Reinforcement Learning
read the original abstract
Distribution Matching Distillation (DMD) facilitates efficient inference by distilling multi-step diffusion models into few-step variants. Concurrently, Reinforcement Learning (RL) has emerged as a vital tool for aligning generative models with human preferences. While both represent critical post-training stages for large-scale diffusion models, existing studies typically treat them as independent, sequential processes, leaving a systematic framework for their unification largely unexplored. In this work, we demonstrate that jointly optimizing these two objectives yields mutual benefits: RL enables more preference-aware and controllable distillation rather than uniformly compressing the full data distribution, while DMD serves as an effective regularizer to mitigate reward hacking during RL training. Building on these insights, we propose DMDR, a unified framework that incorporates Reward-Tilted Distribution Matching optimization alongside two dynamic distillation training strategies in the initial stage, followed by the joint DMD and RL optimization in the second stage. Extensive experiments demonstrate that DMDR achieves state-of-the-art visual quality and prompt adherence among few-step generation methods, even surpassing the performance of its multi-step teacher model.
Forward citations
Cited by 29 Pith papers
-
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.
-
CoDMD: Copula-aware Distribution Matching Distillation for Fast Video Generation
CoDMD adds a copula-matching regularizer to DMD for distilling 50-step video diffusion models to 4 steps, reporting VBench scores of 84.46/84.87 on 1.3B/14B Wan-2.1-T2V models.
-
World Model Self-Distillation: Training World Models to Solve General Tasks
Self-distillation from a caption-conditioned video diffusion model to an image-and-prompt-conditioned executor, enhanced by RL from VLM feedback, enables task solving in world models.
-
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.
-
D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models
D-OPSD formulates supervised fine-tuning of step-distilled diffusion models as on-policy self-distillation by minimizing distribution differences between a text-only student and a multimodal teacher on the student's o...
-
Stream-R1: Reliability-Perplexity Aware Reward Distillation for Streaming Video Generation
Stream-R1 improves distillation of autoregressive streaming video diffusion models by adaptively weighting supervision with a reward model at both rollout and per-pixel levels.
-
Hallo-Live: Real-Time Streaming Joint Audio-Video Avatar Generation with Asynchronous Dual-Stream and Human-Centric Preference Distillation
Hallo-Live achieves 20.38 FPS real-time text-to-audio-video avatar generation with 0.94s latency using asynchronous dual-stream diffusion and HP-DMD preference distillation, matching teacher model quality at 16x highe...
-
Guiding Distribution Matching Distillation with Gradient-Based Reinforcement Learning
GDMD replaces raw-sample rewards with distillation-gradient rewards in RL-guided diffusion distillation, yielding 4-step models that surpass their multi-step teachers on GenEval and human preference metrics.
-
1.x-Distill: Breaking the Diversity, Quality, and Efficiency Barrier in Distribution Matching Distillation
1.x-Distill achieves better quality and diversity than prior few-step distillation methods at 1.67 and 1.74 effective NFEs on SD3 models with up to 33x speedup.
-
Reward Lightning: Fast Video Generation via Homologous Preference Distillation
Homologous preference distillation evaluates adversarial distillation and latent reward alignment on identical latent features, yielding 1–4-step video generators that improve VBench by 2.1% while leading text, motion...
-
Monocular Avatar Reconstruction via Cascaded Diffusion Priors and UV-Space Differentiable Shading
A cascaded LoRA diffusion pipeline in UV space with cross-intrinsic attention and differentiable BRDF shading produces 4K PBR avatar assets from single images after training on under 100 scans.
-
Causal-rCM: A Unified Teacher-Forcing and Self-Forcing Open Recipe for Autoregressive Diffusion Distillation in Streaming Video Generation and Interactive World Models
Causal-rCM unifies teacher-forcing and self-forcing distillation for autoregressive video diffusion, delivering a 2-step model with VBench-T2V score 84.63 and enabling interactive world models on Cosmos 3 using only s...
-
High-Fidelity Two-Step Image Generation via Teacher-Aligned End-to-End Distillation
Z-Image Turbo++ narrows the quality gap to 8-step generation via three distillation techniques tailored for the 2-step regime.
-
Reinforcing Few-step Generators via Reward-Tilted Distribution Matching
RTDMD unifies KL minimization to a reward-tilted teacher into distribution matching plus reward terms, using AC-DMD in stage one and hybrid GRPO-style gradients plus SubGRPO in stage two to reach new SOTA on preferenc...
-
CollectionLoRA: Collecting 50 Effects in 1 LoRA via Multi-Teacher On-Policy Distillation
A multi-teacher distillation framework that packs 50 effect LoRAs and fast sampling into a single adapter while aiming to avoid concept interference.
-
Efficient Image Synthesis with Sphere Latent Encoder
Decouples Sphere Encoder into fixed pretrained encoder and spherical latent denoiser, yielding higher quality and faster inference than the joint original on Animal-Faces, Oxford-Flowers and ImageNet-1K.
-
Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation
Delta Forcing estimates transition consistency from the latent delta between teacher and generator trajectories to balance teacher supervision against a monotonic continuity objective in autoregressive video generation.
-
Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation
Delta Forcing uses latent trajectory deltas to adaptively limit unreliable teacher guidance while enforcing monotonic continuity, improving temporal consistency in interactive autoregressive video generation.
-
Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation
Delta Forcing improves temporal coherence in interactive autoregressive video generation by estimating transition consistency from teacher-generator latent deltas and balancing it against a monotonic continuity objective.
-
FlashMol: High-Quality Molecule Generation in as Few as Four Steps
FlashMol produces chemically valid 3D molecules in 4 steps via distribution matching distillation with respaced timesteps and Jensen-Shannon regularization, matching or exceeding 1000-step teacher performance on QM9 a...
-
D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models
D-OPSD formulates supervised fine-tuning of step-distilled diffusion models as on-policy self-distillation by having the model act as both teacher (with multimodal context) and student (with text-only context) on its ...
-
D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models
D-OPSD enables continuous supervised fine-tuning of few-step diffusion models via on-policy self-distillation where the model acts as both teacher (multimodal context) and student (text-only context) on its own roll-outs.
-
A Test-time Actor-Critic Approach to News Images Generation
A test-time actor-critic approach for refining image generation prompts won the MediaEval NewsImages 2026 challenge.
-
Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models
Lens is a 3.8B-parameter text-to-image model that reaches competitive or superior performance to >6B-parameter systems using 19.3% of the training compute of Z-Image through a densely captioned 800M dataset, multi-res...
-
Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
Z-Image is an efficient 6B-parameter foundation model for image generation that rivals larger commercial systems in photorealism and bilingual text rendering through a new single-stream diffusion transformer and strea...
-
Qwen-Image-Flash: Beyond Objective Design
Empirical analysis of data, guidance, and task mixture in few-step distillation of Qwen-Image-2.0 produces the Qwen-Image-Flash model with improved performance in unified generation and editing tasks.
-
ERNIE-Image Technical Report
The paper presents ERNIE-Image, an open-source 8B DiT text-to-image model claiming leading open-source performance and near-commercial results via specialized data construction and DPO alignment.
-
Delta Forcing: Trust Region Steering for Interactive Autoregressive Video Generation
Delta Forcing constrains unreliable teacher supervision in autoregressive video models using an adaptive trust region based on latent trajectory deltas to improve consistency without losing event responsiveness.
-
Mamoda2.5: Enhancing Unified Multimodal Model with DiT-MoE
Mamoda2.5 is a 25B-parameter DiT-MoE unified AR-Diffusion model that reaches top video generation and editing benchmarks with 4-step inference up to 95.9x faster than baselines.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.