On the Realization of Compositionality in Neural Networks

Joris Baan, Jana Leible, Mitja Nikolaus, David Rau, Dennis Ulmer, Tim Baumgärtner, Dieuwke Hupkes, Elia Bruni


Abstract
We present a detailed comparison of two types of sequence to sequence models trained to conduct a compositional task. The models are architecturally identical at inference time, but differ in the way that they are trained: our baseline model is trained with a task-success signal only, while the other model receives additional supervision on its attention mechanism (Attentive Guidance), which has shown to be an effective method for encouraging more compositional solutions. We first confirm that the models with attentive guidance indeed infer more compositional solutions than the baseline, by training them on the lookup table task presented by Liska et al. (2019). We then do an in-depth analysis of the structural differences between the two model types, focusing in particular on the organisation of the parameter space and the hidden layer activations and find noticeable differences in both these aspects. Guided networks focus more on the components of the input rather than the sequence as a whole and develop small functional groups of neurons with specific purposes that use their gates more selectively. Results from parameter heat maps, component swapping and graph analysis also indicate that guided networks exhibit a more modular structure with a small number of specialized, strongly connected neurons.
Anthology ID:
W19-4814
Volume:
Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP
Month:
August
Year:
2019
Address:
Florence, Italy
Editors:
Tal Linzen, Grzegorz Chrupała, Yonatan Belinkov, Dieuwke Hupkes
Venue:
BlackboxNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
127–137
Language:
URL:
https://aclanthology.org/W19-4814
DOI:
10.18653/v1/W19-4814
Bibkey:
Cite (ACL):
Joris Baan, Jana Leible, Mitja Nikolaus, David Rau, Dennis Ulmer, Tim Baumgärtner, Dieuwke Hupkes, and Elia Bruni. 2019. On the Realization of Compositionality in Neural Networks. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 127–137, Florence, Italy. Association for Computational Linguistics.
Cite (Informal):
On the Realization of Compositionality in Neural Networks (Baan et al., BlackboxNLP 2019)
Copy Citation:
PDF:
https://preview.aclanthology.org/nschneid-patch-5/W19-4814.pdf