AdaWorld: Learning Adaptable World Models with Latent Actions

Shenyuan Gao, Siyuan Zhou, Yilun Du, Jun Zhang, Chuang Gan
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:18744-18771, 2025.

Abstract

World models aim to learn action-controlled future prediction and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this limitation, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-gao25u, title = {{A}da{W}orld: Learning Adaptable World Models with Latent Actions}, author = {Gao, Shenyuan and Zhou, Siyuan and Du, Yilun and Zhang, Jun and Gan, Chuang}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {18744--18771}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/gao25u/gao25u.pdf}, url = {https://proceedings.mlr.press/v267/gao25u.html}, abstract = {World models aim to learn action-controlled future prediction and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this limitation, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.} }
Endnote
%0 Conference Paper %T AdaWorld: Learning Adaptable World Models with Latent Actions %A Shenyuan Gao %A Siyuan Zhou %A Yilun Du %A Jun Zhang %A Chuang Gan %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-gao25u %I PMLR %P 18744--18771 %U https://proceedings.mlr.press/v267/gao25u.html %V 267 %X World models aim to learn action-controlled future prediction and have proven essential for the development of intelligent agents. However, most existing world models rely heavily on substantial action-labeled data and costly training, making it challenging to adapt to novel environments with heterogeneous actions through limited interactions. This limitation can hinder their applicability across broader domains. To overcome this limitation, we propose AdaWorld, an innovative world model learning approach that enables efficient adaptation. The key idea is to incorporate action information during the pretraining of world models. This is achieved by extracting latent actions from videos in a self-supervised manner, capturing the most critical transitions between frames. We then develop an autoregressive world model that conditions on these latent actions. This learning paradigm enables highly adaptable world models, facilitating efficient transfer and learning of new actions even with limited interactions and finetuning. Our comprehensive experiments across multiple environments demonstrate that AdaWorld achieves superior performance in both simulation quality and visual planning.
APA
Gao, S., Zhou, S., Du, Y., Zhang, J. & Gan, C.. (2025). AdaWorld: Learning Adaptable World Models with Latent Actions. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:18744-18771 Available from https://proceedings.mlr.press/v267/gao25u.html.

Related Material