Learning Policies for Multilingual Training of Neural Machine Translation Systems

Low-resource Multilingual Neural Machine Translation (MNMT) is typically tasked with improving the translation performance on one or more language pairs with the aid of highresource language pairs. In this paper, we propose two simple search based curricula – orderings of the multilingual training data – which help improve translation performance in conjunction with existing techniques such as fine-tuning. Additionally, we attempt to learn a curriculum for MNMT from scratch jointly with the training of the translation system with the aid of contextual multi-arm bandits. We show on the FLORES low-resource translation dataset that these learned curricula can provide better starting points for fine tuning and improve overall performance of the translation system.

[1]  Kevin Duh,et al.  Adaptation Data Selection using Neural Language Models: Experiments in Machine Translation , 2013, ACL.

[2]  Ankur Bapna,et al.  Simple, Scalable Adaptation for Neural Machine Translation , 2019, EMNLP.

[3]  Martin Wattenberg,et al.  Google’s Multilingual Neural Machine Translation System: Enabling Zero-Shot Translation , 2016, TACL.

[4]  Orhan Firat,et al.  Massively Multilingual Neural Machine Translation , 2019, NAACL.

[5]  John Langford,et al.  The Epoch-Greedy Algorithm for Multi-armed Bandits with Side Information , 2007, NIPS.

[6]  Processing Systems , 2022, Essentials of Thermal Processing.

[7]  H. Vincent Poor,et al.  Bandit problems with side observations , 2005, IEEE Transactions on Automatic Control.

[8]  Christopher D. Manning,et al.  Stanford Neural Machine Translation Systems for Spoken Language Domains , 2015, IWSLT.

[9]  J. Elman.  Learning and development in neural networks: the importance of starting small , 1993, Cognition.

[10]  Nadir Durrani,et al.  A Deep Fusion Model for Domain Adaptation in Phrase-based MT , 2016, COLING.

[11]  Deniz Yuret,et al.  Transfer Learning for Low-Resource Neural Machine Translation , 2016, EMNLP.

[12]  Huda Khayrallah,et al.  An Empirical Exploration of Curriculum Learning for Neural Machine Translation , 2018, ArXiv.

[13]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[14]  Doina Precup,et al.  Algorithms for multi-armed bandit problems , 2014, ArXiv.

[15]  Marcello Federico,et al.  Improving Zero-Shot Translation of Low-Resource Languages , 2018, IWSLT.

[16]  Ankur Bapna,et al.  Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges , 2019, ArXiv.

[17]  Douglas L. T. Rohde,et al.  Language acquisition in the absence of explicit negative evidence: how important is starting small? , 1999, Cognition.

[18]  Myle Ott,et al.  fairseq: A Fast, Extensible Toolkit for Sequence Modeling , 2019, NAACL.

[19]  William D. Lewis,et al.  Intelligent Selection of Language Model Training Data , 2010, ACL.

[20]  George F. Foster,et al.  Reinforcement Learning based Curriculum Optimization for Neural Machine Translation , 2019, NAACL.

[21]  Kevin Duh,et al.  Curriculum Learning for Domain Adaptation in Neural Machine Translation , 2019, NAACL.

[22]  Markus Freitag,et al.  Fast Domain Adaptation for Neural Machine Translation , 2016, ArXiv.

[23]  Jason Weston,et al.  Curriculum learning , 2009, ICML '09.

[24]  Philipp Koehn,et al.  Two New Evaluation Datasets for Low-Resource Machine Translation: Nepali-English and Sinhala-English , 2019, ArXiv.

[25]  Jianfeng Gao,et al.  Domain Adaptation via Pseudo In-Domain Data Selection , 2011, EMNLP.

[26]  Raj Dabre,et al.  Exploiting Multilingualism through Multistage Fine-Tuning for Low-Resource Neural Machine Translation , 2019, EMNLP.