@inproceedings{3f12c1569a1e41918eec28d77a3a41fc,
title = "Sentence-State LSTMs For Sequence-to-Sequence Learning",
abstract = "Transformer is currently the dominant method for sequence to sequence problems. In contrast, RNNs have become less popular due to the lack of parallelization capabilities and the relatively lower performance. In this paper, we propose to use a parallelizable variant of bi-directional LSTMs (BiLSTMs), namely sentence-state LSTMs (S-LSTM), as an encoder for sequence-to-sequence tasks. The complexity of S-LSTM is only O(n) as compared to O(n2) of Transformer. On four neural machine translation benchmarks, we empirically find that S-SLTM can achieve significantly better performances than BiLSTM and convolutional neural networks (CNNs). When compared to Transformer, our model gives competitive performance while being 1.6 times faster during inference.",
keywords = "Bi-directional LSTMs, CNN, Neural machine translation, Sentence-State LSTMs, Transformers",
author = "Xuefeng Bai and Yafu Li and Zhirui Zhang and Mingzhou Xu and Boxing Chen and Weihua Luo and Derek Wong and Yue Zhang",
note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 10th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2021 ; Conference date: 13-10-2021 Through 17-10-2021",
year = "2021",
doi = "10.1007/978-3-030-88480-2\_9",
language = "英语",
isbn = "9783030884796",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "104--115",
editor = "Lu Wang and Yansong Feng and Yu Hong and Ruifang He",
booktitle = "Natural Language Processing and Chinese Computing - 10th CCF International Conference, NLPCC 2021, Proceedings",
address = "德国",
}