Genesis: Generative scene inference and sampling with object-centric latent representations

M Engelcke, AR Kosiorek, OP Jones… - arxiv preprint arxiv …, 2019 - arxiv.org
Generative latent-variable models are emerging as promising tools in robotics and
reinforcement learning. Yet, even though tasks in these domains typically involve distinct …

Genesis-v2: Inferring unordered object representations without iterative refinement

M Engelcke, O Parker Jones… - Advances in Neural …, 2021 - proceedings.neurips.cc
Advances in unsupervised learning of object-representations have culminated in the
development of a broad range of methods for unsupervised object segmentation and …

Promising or elusive? unsupervised object segmentation from real-world single images

Y Yang, B Yang - Advances in Neural Information …, 2022 - proceedings.neurips.cc
In this paper, we study the problem of unsupervised object segmentation from single
images. We do not introduce a new algorithm, but systematically investigate the …

Unsupervised multi-object segmentation by predicting probable motion patterns

L Karazija, S Choudhury, I Laina… - Advances in …, 2022 - proceedings.neurips.cc
We propose a new approach to learn to segment multiple image objects without manual
supervision. The method can extract objects form still images, but uses videos for …

Compositional scene representation learning via reconstruction: A survey

J Yuan, T Chen, B Li, X Xue - IEEE Transactions on Pattern …, 2023 - ieeexplore.ieee.org
Visual scenes are composed of visual concepts and have the property of combinatorial
explosion. An important reason for humans to efficiently learn from diverse visual scenes is …

Posterior control of blackbox generation

XL Li, AM Rush - arxiv preprint arxiv:2005.04560, 2020 - arxiv.org
Text generation often requires high-precision output that obeys task-specific rules. This fine-
grained control is difficult to enforce with off-the-shelf deep learning models. In this work, we …

Graph ode with factorized prototypes for modeling complicated interacting dynamics

X Luo, Y Gu, H Jiang, J Huang, W Ju, M Zhang… - arxiv preprint arxiv …, 2023 - arxiv.org
This paper studies the problem of modeling interacting dynamical systems, which is critical
for understanding physical dynamics and biological processes. Recent research …

Benchmarking and Analysis of Unsupervised Object Segmentation from Real-World Single Images

Y Yang, B Yang - International Journal of Computer Vision, 2024 - Springer
In this paper, we study the problem of unsupervised object segmentation from single
images. We do not introduce a new algorithm, but systematically investigate the …

Unsupervised video decomposition using spatio-temporal iterative inference

P Zablotskaia, EA Dominici, L Sigal… - arxiv preprint arxiv …, 2020 - arxiv.org
Unsupervised multi-object scene decomposition is a fast-emerging problem in
representation learning. Despite significant progress in static scenes, such models are …

PROVIDE: A probabilistic framework for unsupervised video decomposition

P Zablotskaia, EA Dominici, L Sigal… - Uncertainty in …, 2021 - proceedings.mlr.press
Unsupervised multi-object scene decomposition is a fast-emerging problem in
representation learning. Despite significant progress in static scenes, such models are …