• Computer Science, Mathematics
  • Published in NeurIPS 2019

PHYRE: A New Benchmark for Physical Reasoning

@inproceedings{Bakhtin2019PHYREAN,
  title={PHYRE: A New Benchmark for Physical Reasoning},
  author={Anton Bakhtin and Laurens van der Maaten and Justin Johnson and Laura Gustafson and Ross B. Girshick},
  booktitle={NeurIPS},
  year={2019}
}
Understanding and reasoning about physics is an important ability of intelligent agents. We develop the PHYRE benchmark for physical reasoning that contains a set of simple classical mechanics puzzles in a 2D physical environment. The benchmark is designed to encourage the development of learning algorithms that are sample-efficient and generalize well across puzzles. We test several modern learning algorithms on PHYRE and find that these algorithms fall short in solving the puzzles efficiently… CONTINUE READING

Citations

Publications citing this paper.

References

Publications referenced by this paper.
SHOWING 1-10 OF 60 REFERENCES

Human-level control through deep reinforcement learning

VIEW 7 EXCERPTS
HIGHLY INFLUENTIAL