Learning Physical Intuition of Block Towers by Example

Adam Lerer, Sam Gross, Rob Fergus
Proceedings of The 33rd International Conference on Machine Learning, PMLR 48:430-438, 2016.

Abstract

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 trajectories of the blocks. 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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v48-lerer16, title = {Learning Physical Intuition of Block Towers by Example}, author = {Lerer, Adam and Gross, Sam and Fergus, Rob}, booktitle = {Proceedings of The 33rd International Conference on Machine Learning}, pages = {430--438}, year = {2016}, editor = {Balcan, Maria Florina and Weinberger, Kilian Q.}, volume = {48}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {20--22 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v48/lerer16.pdf}, url = {https://proceedings.mlr.press/v48/lerer16.html}, abstract = {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 trajectories of the blocks. 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.} }
Endnote
%0 Conference Paper %T Learning Physical Intuition of Block Towers by Example %A Adam Lerer %A Sam Gross %A Rob Fergus %B Proceedings of The 33rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Maria Florina Balcan %E Kilian Q. Weinberger %F pmlr-v48-lerer16 %I PMLR %P 430--438 %U https://proceedings.mlr.press/v48/lerer16.html %V 48 %X 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 trajectories of the blocks. 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.
RIS
TY - CPAPER TI - Learning Physical Intuition of Block Towers by Example AU - Adam Lerer AU - Sam Gross AU - Rob Fergus BT - Proceedings of The 33rd International Conference on Machine Learning DA - 2016/06/11 ED - Maria Florina Balcan ED - Kilian Q. Weinberger ID - pmlr-v48-lerer16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 48 SP - 430 EP - 438 L1 - http://proceedings.mlr.press/v48/lerer16.pdf UR - https://proceedings.mlr.press/v48/lerer16.html AB - 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 trajectories of the blocks. 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. ER -
APA
Lerer, A., Gross, S. & Fergus, R.. (2016). Learning Physical Intuition of Block Towers by Example. Proceedings of The 33rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 48:430-438 Available from https://proceedings.mlr.press/v48/lerer16.html.

Related Material