Skip to content
AI News HubLIVE
Original source2 min read

Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching

Summary

This paper presents DM-Align, a single-stage distribution-matching framework that jointly aligns and distills video generative models. By combining a standard distillation gradient with a complementary preference-guided gradient inspired by DPO and GRPO, it avoids the heavy computational cost of reinforcement learning and consistently outperforms standalone and two-stage baselines across multiple foundational video models.

SourcearXiv Computer VisionAuthor: Jiuzhou Lin, Junlong Wu, Fei Zuo, Huan Ouyang, Dewen Fan, Boheng Zhang, Huaiqing Wang, Jia Sun, Fan Yang, Houde Liu, Kehai Chen, Min Zhang, Tingting Gao, Han Li
Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 3 Sep 2026]

Title:Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching

View a PDF of the paper titled Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching, by Jiuzhou Lin and 13 other authors

View PDF HTML (experimental)

Abstract:Aligning video generative models to human preferences heavily relies on Reinforcement Learning (RL), which suffers from extensive computational overhead. Existing workflows typically treat RL and distillation as disconnected stages: applying RL before distillation incurs prohibitive computational costs, whereas applying RL after distillation frequently leads to model collapse. To overcome these limitations, we propose a unified, single-stage optimization framework grounded in Distribution Matching (DM). In the standard DM framework, distillation updates the model via a gradient direction that minimizes the gap between the real and fake models, guiding generations toward clarity and high fidelity. Building upon this, we introduce DM-Align, which derives a complementary gradient direction to guide the model toward human-preferred samples. Inspired by DPO and GRPO, our method leverages the distributional gap -- formulated from either preference pairs or intra-group exploration -- to directly construct this preference-guided gradient. By synergizing these two gradient directions, our approach eliminates the need for multi-step reward evaluation and complex ODE-SDE conversions inherent in traditional RL. Comprehensive experiments across multiple foundational video models demonstrate that this sample-guided framework robustly enhances both distillation quality and preference alignment, consistently outperforming both standalone variants and sequential two-stage pipelines.

Subjects:

Computer Vision and Pattern Recognition (cs.CV)

Cite as: arXiv:2609.04283 [cs.CV]

(or arXiv:2609.04283v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2609.04283

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Junlong Wu [view email] [v1] Thu, 3 Sep 2026 04:46:46 UTC (14,424 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching, by Jiuzhou Lin and 13 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CV

new | recent | 2026-09

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • Existing pipelines treat RL alignment and distillation as separate stages, causing prohibitive compute or model collapse.
  • DM-Align unifies alignment and distillation in one distribution-matching-based optimization stage.
  • A complementary preference-guided gradient is derived from preference pairs or intra-group exploration without multi-step reward evaluation or ODE-SDE conversion.
  • Experiments on multiple fundamental video models show robust gains in both distillation quality and preference alignment.

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models