%0 Generic %1 bowman2015treestructured %A Bowman, Samuel R. %A Manning, Christopher D. %A Potts, Christopher %D 2015 %K ling lstm nn rnn %T Tree-structured composition in neural networks without tree-structured architectures %U http://arxiv.org/abs/1506.04834 %X Tree-structured neural networks encode a particular tree geometry for a sentence in the network design. However, these models have at best only slightly outperformed simpler sequence-based models. We hypothesize that neural sequence models like LSTMs are in fact able to discover and implicitly use recursive compositional structure, at least for tasks with clear cues to that structure in the data. We demonstrate this possibility using an artificial data task for which recursive compositional structure is crucial, and find an LSTM-based sequence model can indeed learn to exploit the underlying tree structure. However, its performance consistently lags behind that of tree models, even on large training sets, suggesting that tree-structured models are more effective at exploiting recursive structure.

@preprint{bowman2015treestructured, abstract = {Tree-structured neural networks encode a particular tree geometry for a sentence in the network design. However, these models have at best only slightly outperformed simpler sequence-based models. We hypothesize that neural sequence models like LSTMs are in fact able to discover and implicitly use recursive compositional structure, at least for tasks with clear cues to that structure in the data. We demonstrate this possibility using an artificial data task for which recursive compositional structure is crucial, and find an LSTM-based sequence model can indeed learn to exploit the underlying tree structure. However, its performance consistently lags behind that of tree models, even on large training sets, suggesting that tree-structured models are more effective at exploiting recursive structure.}, added-at = {2016-10-30T22:42:59.000+0100}, author = {Bowman, Samuel R. and Manning, Christopher D. and Potts, Christopher}, biburl = {https://www.bibsonomy.org/bibtex/2ffdd012eb09bbe32b665d8d772eddd52/jkan}, description = {Tree-structured composition in neural networks without tree-structured architectures}, interhash = {66bff5f6fc4fd15ab4452b608a2660ca}, intrahash = {ffdd012eb09bbe32b665d8d772eddd52}, keywords = {ling lstm nn rnn}, note = {cite arxiv:1506.04834Comment: To appear in the proceedings of the 2015 NIPS Workshop on Cognitive Computation: Integrating Neural and Symbolic Approaches}, timestamp = {2016-10-30T22:42:59.000+0100}, title = {Tree-structured composition in neural networks without tree-structured architectures}, url = {http://arxiv.org/abs/1506.04834}, year = 2015 }