analiza aiqbench
This paper introduces 4DGS-JEPA, a novel architecture that shifts dynamic Gaussian Splatting from pure reconstruction toward learning reusable predictive dynamics for evolving 3D scenes. By leveraging Joint-Embedding Predictive Architectures (JEPA) principles, the model achieves causal multi-horizon prediction through temporal composition without requiring complete future appearance reconstruction. The approach theoretically characterizes zero-loss path agreement and demonstrates robustness against topology changes via a hybrid correspondence mechanism, offering a significant advancement in how 3D worlds are modeled causally.
4DGS-JEPA: Temporally Compositional Joint-Embedding Prediction for Dynamic Gaussian Splatting
arXiv:2609.25036v1 Announce Type: new Abstract: Dynamic Gaussian Splatting provides an explicit representation of evolving 3D scenes, but existing approaches are primarily optimized for reconstruction, future-state generation, or rendering rather than for learning reusable predictive dynamics. We propose 4DGS-JEPA, a Gaussian-native joint-embedding predictive architecture for causal multi-horizon prediction over dynamic Gaussian scenes. The model uses a hierarchical scene-, motion-group-, and Gaussian-level representation together with a horizon-conditioned transition operator that supports both direct prediction and recursive rollout. Its central principle is temporal composition: different chronological transition paths reaching the same future endpoint should produce compatible predictive states. Endpoint and multi-horizon path supervision anchor these predictions to future target embeddings, while a selective geometry decoder and geometry-level composition ground the learned dynamics in consistent group motion and Gaussian geometry without requiring complete future appearance reconstruction. We further introduce a hybrid correspondence mechanism that combines persistent canonical identity with residual optimal-transport matching under reordering and topology change. We characterize zero-loss path agreement and finite-error rollout accumulation theoretically. Three controlled experiments provide mechanism-level evidence that temporal composition reduces latent path dependence while retaining predictive accuracy, geometry-level composition improves consistency of decoded motion, and hybrid correspondence preserves reliable identity while remaining robust when correspondence becomes ambiguous. Together, 4DGS-JEPA provides a predictive, temporally compositional formulation of dynamic Gaussian worlds.
Abstract:Dynamic Gaussian Splatting provides an explicit representation of evolving 3D scenes, but existing approaches are primarily optimized for reconstruction, future-state generation, or rendering rather than for learning reusable predictive dynamics. We propose 4DGS-JEPA, a Gaussian-native joint-embedding predictive architecture for causal multi-horizon prediction over dynamic Gaussian scenes. The model uses a hierarchical scene-, motion-group-, and Gaussian-level representation together with a horizon-conditioned transition operator that supports both direct prediction and recursive rollout. Its central principle is temporal composition: different chronological transition paths reaching the same future endpoint should produce compatible predictive states. Endpoint and multi-horizon path supervision anchor these predictions to future target embeddings, while a selective geometry decoder and geometry-level composition ground the learned dynamics in consistent group motion and Gaussian geometry without requiring complete future appearance reconstruction. We further introduce a hybrid correspondence mechanism that combines persistent canonical identity with residual optimal-transport matching under reordering and topology change. We characterize zero-loss path agreement and finite-error rollout accumulation theoretically. Three controlled experiments provide mechanism-level evidence that temporal composition reduces latent path dependence while retaining predictive accuracy, geometry-level composition improves consistency of decoded motion, and hybrid correspondence preserves reliable identity while remaining robust when correspondence becomes ambiguous. Together, 4DGS-JEPA provides a predictive, temporally compositional formulation of dynamic Gaussian worlds.
Artificial Intelligence (cs.AI)
arXiv:2609.25036 [cs.AI]
(or arXiv:2609.25036v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.25036 Focus to learn more arXiv-issued DOI via DataCite
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.
Text integral sub licența deschisă a publicației · 299 cuvinte
Subiecte conexe
MDN-VJEPA: Multimodal Probabilistic Joint-Embedding World Models
arXiv · 14 oct. 2026
Your Probabilistic JEPA Is Secretly a Hidden Markov Model: A State-Space Interpretation of Joint-Embedding Predictive Learning
arXiv · 17 aug. 2026
Scalable and Efficient Joint Spiking Embedding Predictive Architecture for Large-Scale Dynamic Graphs
arXiv · 22 iul. 2026
Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization
arXiv · 11 sept. 2026