Fetching the paper…

Reading the bibliography…

2020

Solving Physics Puzzles by Reasoning about Paths

Harter, Augustin, Melnik, Andrew, Kumar, Gaurav et al.

Original
PDF
Post on X

Understand

We propose a new deep learning model for goal-driven tasks that require intuitive physical reasoning and intervention in the scene to achieve a desired end goal.

Its modular structure is motivated by hypothesizing a sequence of intuitive steps that humans apply when trying to solve such a task.
The model first predicts the path the target object would follow without intervention and the path the target object should follow in order to solve the task.
Next, it predicts the desired path of the action object and generates the placement of the action object.

Solving Physics Puzzles by Reasoning about Paths · Around

Built on

Learning plannable representations with causal infogan
Thanard Kurutach, Aviv Tamar, Ge Yang, Stuart J Russell, and Pieter Abbeel · 2018
Earlier work this paper cites.
Modularization of end-to-end learning: Case study in arcade games
Original
Andrew Melnik, Sascha Fleer, Malte Schilling, and Helge Ritter · 2018
Earlier work this paper cites.
Phyre: A new benchmark for physical reasoning
Anton Bakhtin, Laurens van der Maaten, Justin Johnson, Laura Gustafson, and Ross Girshick · 2019
Earlier work this paper cites.

Similar

Jointly trained variational autoencoder for multi-modal sensor fusion
Timo Korthals, Marc Hesse, Jürgen Leitner, Andrew Melnik, and Ulrich Rückert · 2019
Cited alongside, same era.
The tools challenge: Rapid trial-and-error learning in physical problem solving
Original
Kelsey R Allen, Kevin A Smith, and Joshua B Tenenbaum · 2019
Cited alongside, same era.
Forward prediction for physical reasoning
Original
Rohit Girdhar, Laura Gustafson, Aaron Adcock, and Laurens van der Maaten · 2020
Cited alongside, same era.

Then

An error-based addressing architecture for dynamic model learning
Nicolas Bach, Andrew Melnik, Federico Rosetto, and Helge Ritter · 2020
Closest in time.
Learn to move through a combination of policy gradient algorithms: Ddpg, d4pg, and td3
Nicolas Bach, Andrew Melnik, Malte Schilling, Timo Korthals, and Helge Ritter · 2020
Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…