Generalization without Systematicity: On the Compositional Skills of Sequence-to-Sequence Recurrent Networks

@inproceedings{Lake2018GeneralizationWS,
  title={Generalization without Systematicity: On the Compositional Skills of Sequence-to-Sequence Recurrent Networks},
  author={Brenden M. Lake and Marco Baroni},
  booktitle={ICML},
  year={2018}
}
Humans can understand and produce new utterances effortlessly, thanks to their compositional skills. Once a person learns the meaning of a new verb “dax,” he or she can immediately understand the meaning of “dax twice” or “sing and dax.” In this paper, we introduce the SCAN domain, consisting of a set of simple compositional navigation commands paired with the corresponding action sequences. We then test the zero-shot generalization capabilities of a variety of recurrent neural networks (RNNs… CONTINUE READING
Recent Discussions
This paper has been referenced on Twitter 104 times over the past 90 days. VIEW TWEETS

Similar Papers

Loading similar papers…