Strongly-Typed Recurrent Neural Networks

@inproceedings{Balduzzi2016StronglyTypedRN,
  title={Strongly-Typed Recurrent Neural Networks},
  author={David Balduzzi and Muhammad Ghifary},
  booktitle={ICML},
  year={2016}
}
Recurrent neural networks are increasing popular models for sequential learning. Unfortunately, although the most effective RNN architectures are perhaps excessively complicated, extensive searches have not found simpler alternatives. This paper imports ideas from physics and functional programming into RNN design to provide guiding principles. From physics, we introduce type constraints, analogous to the constraints that forbids adding meters to seconds. From functional programming, we require… CONTINUE READING
Highly Cited
This paper has 23 citations. REVIEW CITATIONS
Recent Discussions
This paper has been referenced on Twitter 42 times over the past 90 days. VIEW TWEETS

References

Publications referenced by this paper.
Showing 1-10 of 29 references

Similar Papers

Loading similar papers…