Generalization without systematicity: On the compositional skills of sequence-to-sequence recurrent networks

Abstract

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) trained on SCAN with sequence-to-sequence methods. We find that RNNs can make successful zero-shot generalizations when the differences between training and test commands are small, so that they can apply "mix-and-match" strategies to solve the task. However, when generalization requires systematic compositional skills (as in the "dax" example above), RNNs fail spectacularly. We conclude with a proof-of-concept experiment in neural machine translation, suggesting that lack of systematicity might be partially responsible for neural networks' notorious training data thirst.

Published at the 35th International Conference on Machine Learning (ICML 2018)

Cited by in corpus (6)

• Compositional generalization in a deep seq2seq model by separating syntax and semantics
• Enhancing the Transformer with Explicit Relational Encoding for Math Problem Solving
• Transforming task representations to perform novel tasks
• Visually Grounded Continual Learning of Compositional Phrases
• The relational processing limits of classic and contemporary neural network models of language processing
• Recursive Top-Down Production for Sentence Generation with Latent Trees