While recurrent neural networks have found success in a variety of natural language processing applications, they are general models of sequential data. We investigate how the properties of natural language data affect an LSTM's ability to learn a nonlinguistic task: recalling elements from its input. We find that models trained on natural language data are able to recall tokens from much longer sequences than models trained on non-language sequential data. Furthermore, we show that the LSTM learns to solve the memorization task by explicitly using a subset of its neurons to count timesteps in the input. We hypothesize that the patterns and structure in natural language data enable LSTMs to learn by providing approximate ways of reducing loss, but understanding the effect of different training data on the learnability of LSTMs remains an open question.
|Original language||American English|
|Title of host publication||ACL 2018 - Representation Learning for NLP, Proceedings of the 3rd Workshop|
|Publisher||Association for Computational Linguistics (ACL)|
|Number of pages||7|
|State||Published - 2018|
|Event||3rd Workshop on Representation Learning for NLP, RepL4NLP 2018 at the 56th Annual Meeting of the Association for Computational Linguistics ACL 2018 - Melbourne, Australia|
Duration: 20 Jul 2018 → …
|Name||Proceedings of the Annual Meeting of the Association for Computational Linguistics|
|Conference||3rd Workshop on Representation Learning for NLP, RepL4NLP 2018 at the 56th Annual Meeting of the Association for Computational Linguistics ACL 2018|
|Period||20/07/18 → …|
Bibliographical noteFunding Information:
We thank the ARK as well as the anonymous reviewers for their valuable feedback. NL is supported by a Washington Research Foundation Fellowship and a Barry M. Goldwater Scholarship. This work was supported in part by a hardware gift from NVIDIA Corporation and a UW High Performance Computing Club Cloud Credit Award.
© 2018 Association for Computational Linguistics.