Compositional Latent Dynamics for Long-Horizon World Modelling

Abstract

We introduce CLD, an architecture in which latent dynamics are composed from a discrete library of learned operators rather than predicted by a monolithic transition model. CLD trained on 14,000 hours of multimodal interaction data extrapolates physical dynamics across regimes the monolithic baseline never sees (impact, contact-rich manipulation, two-body wave interaction) and produces counterfactual rollouts whose final-state error is 2.4× lower than the previous state of the art at 200-step horizons.

Index metadata

Cell

voronoi-19, bessel-04

Compute

211 H100-days + 18 TPU-v5-pod-hours

Status

Open release

Code

github.com/alphabell-labs/cld

Data

subset released; full set under sovereign partner licence

DOI

10.48550/arXiv.2508.58450

arXiv

arXiv:2508.58450

What this paper is part of

This index entry is part of the World models research axis. The producing cell — voronoi-19 — collaborates with adjacent cells listed in the cell directory. The paired interpretability cell (where applicable) is identified in the metadata above; their disagreement reports — if any — accompany the public release.

How to read this

If you want to use the result: the code (where available) is at https://github.com/alphabell-labs/ab-composit; the dataset is at TBD when one is released. To cite this report, prefer the DOI/arXiv identifier and the BibTeX block above. To discuss this with the producing cell, contact the lab with the index entry slug compositional-latent-dynamics.

Limitations

Each cell-published report carries an explicit limitations section in the internal index. We do not paraphrase it here. Read the linked PDF — particularly its limitations and threats-to-validity sections — before downstream use.