World models are a fundamental component in model-based reinforcement
le...
Despite remarkable recent advances, making object-centric learning work ...
Unsupervised object-centric representation (OCR) learning has recently d...
In this paper, we propose a novel object-centric representation, called
...
Unsupervised object-centric learning aims to represent the modular,
comp...
Top-performing Model-Based Reinforcement Learning (MBRL) agents, such as...
Transformers have been successful for many natural language processing t...
Object-centric world models provide structured representation of the sce...
Reconciling symbolic and distributed representations is a crucial challe...
The remarkable recent advances in object-centric generative world models...
When tasks change over time, meta-transfer learning seeks to improve the...
A crucial ability of human intelligence is to build up models of individ...
The ability to decompose complex multi-object scenes into meaningful
abs...
Compositional structures between parts and objects are inherent in natur...
The main limitation of previous approaches to unsupervised sequential
ob...
For embodied agents to infer representations of the underlying 3D physic...
We introduce a variational approach to learning and inference of tempora...
Neural processes combine the strengths of neural networks and Gaussian
p...
Imitation learning is an effective alternative approach to learn a polic...
Single-view 3D shape reconstruction is an important but challenging prob...
Learning to infer Bayesian posterior from a few-shot dataset is an impor...
Current language models have a significant limitation in the ability to
...
Memory networks are neural networks with an explicit memory component th...
The problem of rare and unknown words is an important issue that can
pot...
Over the past decade, large-scale supervised learning corpora have enabl...
We propose a stochastic gradient Markov chain Monte Carlo (SG-MCMC) algo...
Despite having various attractive qualities such as high prediction accu...
In this paper we address the following question: Can we approximately sa...