Corpus ID: 88524238

Out-of-the box neural networks can support combinatorial generalization

@article{Vankov2019OutoftheBN,
  title={Out-of-the box neural networks can support combinatorial generalization},
  author={Ivan Vankov and Jeffrey S. Bowers},
  journal={ArXiv},
  year={2019},
  volume={abs/1903.12354}
}
  • Ivan Vankov, Jeffrey S. Bowers
  • Published 2019
  • Computer Science
  • ArXiv
  • Combinatorial generalization - the ability to understand and produce novel combinations of already familiar elements - is considered to be a core capacity of the human mind and a major challenge to neural network models. A significant body of research suggests that conventional neural networks can't solve this problem unless they are endowed with mechanisms specifically engineered for the purpose of representing symbols. In this paper we introduce a novel way of representing symbolic structures… CONTINUE READING

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 36 REFERENCES

    Getting symbols out of a neural architecture

    VIEW 1 EXCERPT

    Rethinking Eliminative Connectionism

    VIEW 1 EXCERPT