Learning Physical Intuition of Block Towers by Example
This work addresses the challenge of learning intuitive physics for AI systems, which is incremental in applying existing deep learning methods to a specific domain.
The paper tackled the problem of teaching deep convolutional networks to predict the stability and collapse trajectories of block towers using synthetic 3D data, achieving accurate predictions and generalization to new scenarios and real images with performance comparable to humans.
Wooden blocks are a common toy for infants, allowing them to develop motor skills and gain intuition about the physical behavior of the world. In this paper, we explore the ability of deep feed-forward models to learn such intuitive physics. Using a 3D game engine, we create small towers of wooden blocks whose stability is randomized and render them collapsing (or remaining upright). This data allows us to train large convolutional network models which can accurately predict the outcome, as well as estimating the block trajectories. The models are also able to generalize in two important ways: (i) to new physical scenarios, e.g. towers with an additional block and (ii) to images of real wooden blocks, where it obtains a performance comparable to human subjects.