arc

ARC can be seen as a general artificial intelligence benchmark, as a program synthesis benchmark, or as a psychometric intelligence test. It is targeted at both humans and artificially intelligent systems that aim at emulating a human-like form of general fluid intelligence.

• URL: https://github.com/fchollet/ARC/
• DatasetBuilder: tfds.image.arc.ARC

arc is configured with tfds.image.arc.ARCConfig and has the following configurations predefined (defaults to the first one):

• 2019-12-06 (v1.0.0) (Size: Unknown size): ARC commit bd9e2c9 from 2019-12-06

arc/2019-12-06

ARC commit bd9e2c9 from 2019-12-06

Versions:

• 1.0.0 (default):

Statistics

Split	Examples
ALL	800
TEST	400
TRAIN	400

Features

FeaturesDict({
    'task_id': Text(shape=(), dtype=tf.string),
    'test': Sequence({
        'input': Sequence(Sequence(Tensor(shape=(), dtype=tf.int32))),
        'output': Sequence(Sequence(Tensor(shape=(), dtype=tf.int32))),
    }),
    'train': Sequence({
        'input': Sequence(Sequence(Tensor(shape=(), dtype=tf.int32))),
        'output': Sequence(Sequence(Tensor(shape=(), dtype=tf.int32))),
    }),
})

Homepage

• https://github.com/fchollet/ARC/

Citation

@misc{chollet_francois_2019,
  title     = {The Measure of Intelligence},
  url       = {https://arxiv.org/abs/1911.01547},
  journal   = {arXiv.org},
  author    = {Francois Chollet},
  year      = {2019},
  month     = {Nov}
}

---