thunlp-mt / mt-reading-list Goto Github PK

@inproceedings{Feng2016Improving,
author = {Shi Feng and
Shujie Liu and
Nan Yang and
Mu Li and
Ming Zhou and
Kenny Q. Zhu},
title = {Improving Attention Modeling with Implicit Distortion and Fertility for Machine Translation},
booktitle = {{COLING} 2016, 26th International Conference on Computational Linguistics,
Proceedings of the Conference: Technical Papers, December 11-16, 2016,
Osaka, Japan},
pages = {3082--3092},
year = {2016},
}

Particularly, I'm looking for papers relating to incorporating domain glossaries and improving accuracy/consistency of number translations in neural machine translation

Hi there,

Here is another paper about document-level translation (which can also deal with single-sentence translation):

Zaixiang Zheng, Xiang Yue, Shujian Huang, Jiajun Chen, Alexandra Birch. 2020. Towards Making the Most of Context in Neural Machine Translation. In Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence (IJCAI).

ijcai version | arxiv version (code available)

Many thanks!
Zaixiang

• ACL 2014
• PDF link: https://aclanthology.info/papers/P14-2046/p14-2046
• MT, this paper develops a system that lets people overcome language barriers by letting them
  speak a language they do not know. That system accepts text entered by a user,
  translates the text, then converts the translation into a phonetic spelling in the user’s
  own orthography.
• For example: input:"interesting"-->output:"因吹斯听"

Yiming Wang, Fei Tian, Dongjian He, Tao Qin, ChengXiang Zhai, Tie-Yan Liu. 2019. Non-Autoregressive Machine Translation with Auxiliary Regularization. In Proceedings of AAAI 2019.

The first and third authors have wrong names.

----------------------------------->

Yiren Wang, Fei Tian, Di He, Tao Qin, ChengXiang Zhai, Tie-Yan Liu. 2019. Non-Autoregressive Machine Translation with Auxiliary Regularization. In Proceedings of AAAI 2019.

• arxiv(cs.cl) 2016
• PDF link: https://arxiv.org/pdf/1610.10099v1.pdf
• MT & Language Modeling, this paper introduced 'ByteNet', a character level dilated conv NN based encoder-decoder model, which encouraged a line of research (e.g., Transformer) and achieved two inspiring and insightful results in that time:

* The ByteNet decoder attains state-of-the-art performance on character-level 
language modelling and outperforms the previous best results obtained with 
recurrent neural networks. 

* The ByteNet also achieves performance on raw character-level machine 
translation that approaches that of the best neural translation models that 
run in quadratic time.

ACL 2018
pdf: http://aclweb.org/anthology/P18-1073
intro: This paper introduced a robust self-learning method to learn an unsupervised bilingual word mapping, and use it to induce bilingual lexicons. They claimed state-of-the-art results on the dataset of Dinu et al. (2015) and the extensions of Artetxe et al. (2017, 2018a)

@inproceedings{xia2017dualsupervised,
title={Dual Supervised Learning.},
author={Xia, Yingce and Qin, Tao and Chen, Wei and Bian, Jiang and Yu, Nenghai and Liu, Tieyan},
journal={international conference on machine learning},
pages={3789--3798},
year={2017}}

@inproceedings{Xia2017DualInference,
author = {Yingce Xia and
Jiang Bian and
Tao Qin and
Nenghai Yu and
Tie{-}Yan Liu},
title = {Dual Inference for Machine Learning},
booktitle = {Proceedings of the Twenty-Sixth International Joint Conference on
Artificial Intelligence, {IJCAI} 2017, Melbourne, Australia, August
19-25, 2017},
pages = {3112--3118},
year = {2017},
crossref = {DBLP:conf/ijcai/2017},
url = {https://doi.org/10.24963/ijcai.2017/434},
doi = {10.24963/ijcai.2017/434},
timestamp = {Wed, 27 Jun 2018 12:24:11 +0200},
biburl = {https://dblp.org/rec/bib/conf/ijcai/XiaBQYL17},
bibsource = {dblp computer science bibliography, https://dblp.org}
}

Fairness and Diversity

Tianxiao Shen, Myle Ott, Michael Auli, Marc'Aurelio Ranzato:
Mixture Models for Diverse Machine Translation: Tricks of the Trade. ICML 2019: 5719-5728

http://proceedings.mlr.press/v97/shen19c.html

"Shumin Shi" should be "Shuming Shi"

Zhaopeng Tu, Yang Liu, Zhengdong Lu, Xiaohua Liu, and Hang Li. 2017. Context Gates for Neural Machine Translation. Transactions of the Association for Computational Linguistics. (Citation: 36)

This paper is essentially about how to balance source-side and target-side context in sentence-level MT. The paper might be inappropriately categorized into "document-level translation".

I personally suggest it could be put into "Coverage Constraints".

Recently, there are more and more works in NAT area, I am wondering if it is necessary to create a new sub-topic?

Accepted by AAAI 2020
https://arxiv.org/abs/1911.09333
A paper aiming at enhancing the diversity of NMT by taking advantage of multi-head attention.

Hello,
Very nice list of papers covering a range of topics related to modern MT techniques.
I thought the below paper would be a good addition to the multilingual MT models.

Parameter Sharing Methods for Multilingual Self-Attentional Translation Models
http://aclweb.org/anthology/W18-6327

Best,

Thank you for your awesome MT-Reading-List. I suggest adding algorithms used in top performance WMT systems, because some papers are just papers which are not effective when data are abundant. Furthermore, an ensemble Transformer + BPE + Back-translation is a strong baseline in practice. The algorithms employed in WMT competitions will clarify which idea actually works when data are abundant.

Actually the link provided for this paper is wrong, and here is the correct link:
https://www.aclweb.org/anthology/2020.acl-main.37.pdf

• ICLR 2019 from OpenReview
• PDF link: https://openreview.net/pdf?id=ByecAoAqK7
• OpenReview page: https://openreview.net/forum?id=ByecAoAqK7
• TL;DR: A multilingual NMT model with reinforcement learning (dual learning) aiming to improve zero-shot translation directions.
• Recommended by @alphadl. Thanks!

The link leading to Adam Lopez's paper Statistical Machine Translation is not working.

SMT > Tutorials > Adam Lopez. 2008. Statistical Machine Translation

@inproceedings{tu2017neural,
title={Neural machine translation with reconstruction},
author={Tu, Zhaopeng and Liu, Yang and Shang, Lifeng and Liu, Xiaohua and Li, Hang},
booktitle={Thirty-First AAAI Conference on Artificial Intelligence},
year={2017}
}

• We update papers about machine translation from top conferences, including ICLR, NeurIPS, ICML, ACL, EMNLP, NAACL, COLING, EACL and so on, as well as top journals including CL and TACL.
• Currently, we only add officially published papers as well as the archived papers which have triggered heated discussion (BERT, for example!).
• For those recently archived papers, or insightful but accidentally rejected papers, we tend to follow them by opening issues; we'll close the issues as soon as the corresponding papers are accepted. Discussions about them are also welcomed~
• Last but not least, feel free to recommend more papers!

Hi,

I really appreciate your hard work which facilitates the literature review in MT. I wonder if one of my works can be added to the list :)

ACL 2021 main conference
PDF link: On Compositional Generalization of Neural Machine Translation
TL;DR: Quantitative and systematic analysis of compositional generalization in NMT with a new testbed for related research in future.

Hi there! I open this issue to suggest this EMNLP'19 paper Dynamic past and future for neural machine translation, which proposes a guided dynamic routing mechanism upon capsule networks to distinguish translated and untranslated contents during translation.

Thx!
zaixiang