Mahi Luthra, Jiayi Shen, Maxime Poli, Angelo Ortiz Tandazo, Yosuke Higuchi, Youssef Benchekroun, Martin Gleize, Charles-Éric Saint-James, Dongyan Lin, Phillip Rust, Angel Villar-Corrales, Surya, Vanessa Stark, Rashel Moritz, Juan Pino, Yann LeCun, Emmanuel Dupoux. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Learning a latent dynamics model provides a task-agnostic representation of an agent's understanding of its environment. Leveraging this knowledge for model-based reinforcement learning (RL) holds the potential to improve sample efficiency over model-free methods by learning from imagined rollouts. Furthermore, because the latent space serves as input to behavior models, the informative representations learned by the world model facilitate efficient learning of desired skills. Most existing methods rely on holistic representations of the environment’s state. In contrast, humans reason about objects and their interactions, predicting how actions will affect specific parts of their surroundings. Inspired by this, we propose *Slot-Attention for Object-centric Latent Dynamics (SOLD)*, a novel model-based RL algorithm that learns object-centric dynamics models in an unsupervised manner from pixel inputs. We demonstrate that the structured latent space not only improves model interpretability but also provides a valuable input space for behavior models to reason over. Our results show that SOLD outperforms DreamerV3 and TD-MPC2 - state-of-the-art model-based RL algorithms - across a range of multi-object manipulation environments that require both relational reasoning and dexterous control. Videos and code are available at https:// slot-latent-dynamics.github.io.

Predicting future scene representations is a crucial task for enabling robots to understand and interact with the environment. However, most existing methods rely on videos and simulations with precise action annotations, limiting their ability to leverage the large amount of available unlabeled video data. To address this challenge, we propose PlaySlot, an object-centric video prediction model that infers object representations and latent actions from unlabeled video sequences. It then uses these representations to forecast future object states and video frames. PlaySlot allows the generation of multiple possible futures conditioned on latent actions, which can be inferred from video dynamics, provided by a user, or generated by a learned action policy, thus enabling versatile and interpretable world modeling. Our results show that PlaySlot outperforms both stochastic and object-centric baselines for video prediction across different environments. Furthermore, we show that our inferred latent actions can be used to learn robot behaviors sample-efficiently from unlabeled video demonstrations. Videos and code are available on our project website.

We present a framework for object-centric video prediction, i.e., parsing a video sequence into objects, and modeling their dynamics and interactions in order to predict the future object states from which video frames are rendered. To facilitate the learning of meaningful spatio-temporal object representations and forecasting of their states, we propose two novel object-centric video prediction (OCVP) transformer modules, which decouple the processing of temporal dynamics and object interactions. We show how OCVP predictors outperform object-agnostic video prediction models on two different datasets. Furthermore, we observe that OCVP modules learn consistent and interpretable object representations. Animations and code to reproduce our results can be found in our project website1.

Recent advances in deep learning have led to significant improvements in single image super-resolution (SR) research. However, due to the amplification of noise during the upsampling steps, state-of-the-art methods often fail at reconstructing high-resolution images from noisy versions of their low-resolution counterparts. However, this is especially important for images from unknown cameras with unseen types of image degradation. In this work, we propose to jointly perform denoising and super-resolution. To this end, we investigate two architectural designs: "in-network" combines both tasks at feature level, while "pre-network" first performs denoising and then super-resolution. Our experiments show that both variants have specific advantages: The in-network design obtains the strongest results when the type of image corruption is aligned in the training and testing dataset, for any choice of denoiser. The pre-network design exhibits superior performance on unseen types of image corruption, which is a pathological failure case of existing super-resolution models. We hope that these findings help to enable super-resolution also in less constrained scenarios where source camera or imaging conditions are not well controlled. Source code and pretrained models are available at https://github.com/angelvillar96/super-resolution-noisy-images.