Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence

Hugging Face Daily Papers Papers

Summary

LingBot-Video presents a DiT-based video pretraining framework with Mixture-of-Experts architecture, specialized data augmentation, and multi-dimensional reward system for embodied intelligence applications.

Despite the recent promise in robot control, video generative models suffer from a domain mismatch due to their primary focus on content creation. For example, their design inherently prioritizes visual fidelity and creativity over computational efficiency and physical realism. In this work, we present LingBot-Video, a DiT-based video pretraining paradigm specifically tailored for embodied intelligence. From the architecture perspective, we adopt the Mixture-of-Experts (MoE), instead of dense, framework to achieve a better trade-off between modeling capacity and inference efficiency, and manage to scale it up from scratch. From the data perspective, we construct a data profiling engine that augments standard internet videos with extensive robot-oriented footage, encompassing manipulation, navigation, and egocentric perspectives, to equip the base model with an intrinsic understanding of actions and world dynamics. From the training perspective, we develop a multi-dimensional reward system to enforce the alignment regarding physical rationality and task completion, going beyond standard criteria such as aesthetics, prompt-following, and motion consistency. Comprehensive evaluations validate its performance and efficiency as a video foundation model. We contribute LingBot-Video as the inaugural large-scale, open-source MoE video foundation model to the community, in a pioneering effort to bridge digital creativity and physical actuation.
Original Article
View Cached Full Text

Cached at: 07/09/26, 07:51 AM

Paper page - Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence

Source: https://huggingface.co/papers/2607.07675 Published on Jul 8

#2 Paper of the day Authors:

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

,

Abstract

LingBot-Video presents a DiT-based video pretraining framework with Mixture-of-Experts architecture, specialized data augmentation, and multi-dimensional reward system for embodied intelligence applications.

Despite the recent promise in robot control, video generative models suffer from a domain mismatch due to their primary focus on content creation. For example, their design inherently prioritizes visual fidelity and creativity over computational efficiency and physical realism. In this work, we present LingBot-Video, aDiT-based video pretrainingparadigm specifically tailored forembodied intelligence. From the architecture perspective, we adopt theMixture-of-Experts(MoE), instead of dense, framework to achieve a better trade-off between modeling capacity and inference efficiency, and manage to scale it up from scratch. From the data perspective, we construct adata profiling enginethat augments standard internet videos with extensiverobot-oriented footage, encompassing manipulation, navigation, and egocentric perspectives, to equip the base model with an intrinsic understanding of actions and world dynamics. From the training perspective, we develop amulti-dimensional reward systemto enforce the alignment regardingphysical rationalityandtask completion, going beyond standard criteria such as aesthetics, prompt-following, and motion consistency. Comprehensive evaluations validate its performance and efficiency as a video foundation model. We contribute LingBot-Video as the inaugural large-scale, open-source MoE video foundation model to the community, in a pioneering effort to bridge digital creativity and physical actuation.

View arXiv pageView PDFProject pageGitHub77Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.07675 in a model README.md to link it from this page.

Datasets citing this paper1

#### cy0307/awesome-egocentric-atlas Viewer• Updatedabout 1 hour ago • 816 • 1.21k • 5

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.07675 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

EMO: Pretraining Mixture of Experts for Emergent Modularity

Hugging Face Daily Papers

EMO is a Mixture-of-Experts model that enables modular deployment by grouping similar domain tokens with shared experts, achieving performance comparable to standard MoEs while allowing significant expert pruning (25% experts retain 99% performance) without performance degradation.