Learning to see physics via visual de-animation
We introduce a paradigm for understanding physical scenes without human annotations. At the core of our system is a physical world representation that is first recovered by a perception module and then utilized by physics and graphics engines. During training, the perception module and the generativ...
Main Authors: | Wu, Jiajun (Author), Lu, Erika (Author), Kohli, Pushmeet (Author), Freeman, William T (Author), Tenenbaum, Joshua B (Author) |
---|---|
Other Authors: | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory (Contributor) |
Format: | Article |
Language: | English |
Published: |
Neural Information Processing Systems Foundation, Inc,
2021-02-09T21:08:02Z.
|
Subjects: | |
Online Access: | Get fulltext |
Similar Items
-
Neural scene de-rendering
by: Wu, Jiajun, et al.
Published: (2020) -
Learning to see the physical world
by: Wu, Jiajun,Ph.D.Massachusetts Institute of Technology.
Published: (2020) -
The neuro-symbolic concept learner: Interpreting scenes, words, and sentences from natural supervision
by: Mao, Jiayuan, et al.
Published: (2020) -
Galileo: Perceiving physical object properties by integrating a physics engine with deep learning
by: Yildirim, Ilker, et al.
Published: (2018) -
Learning a probabilistic latent space of object shapes via 3D generative-adversarial modeling
by: Wu, Jiajun, et al.
Published: (2017)