Welcome to Wen Lai's Blog

In our previous paper published in COLING 2022, we investigate the domain robustness and domain adaptability in machine translation using meta-learning. As an extension of our COLING 2022 paper, we investigate the methods in multilingual scenarios, which adapting the multilingual neural machien translation (MNMT) model to both a new domain and t... Read More

In previous post, we introduced the meta-learning technology used in machine translation. In this post, we prepare to introduce our paper Improving Both Domain Robustness and Domain Adaptability in Machine Translation published in COLING 2022 more details. Background and Motivation The success of Neural Machine Translation (NMT) heavily relies... Read More

Meta-learning, also known as “learning to learn”, has been shown to allow faster finetuning, converge to better performance, and achieve outstanding results for few-shot learning in many applications. It is believed that meta-learning has excellent potential to be applied in NLP, and some works has been proposed with notable achievements in sev... Read More

In this post, I would like to introduce a tutorial in NAACL 2022 named Contrastive Data and Learning for Natural Language Processing. The tutorial introduce some recent works in NLP using contrastive learning techniques. Fore more details, I’d recommend you refer to the tutorials website and the paper list of contrastive learning. In addition, I... Read More

In this post, I would like to introduce a survey paper titled Domain Adaptation and Multi-Domain Adaptation for Neural Machine Translation: A Survey. Here, I also recommend the readers to read another survey paper of domain adaptation in NMT A Survey of Domain Adaptation for Neural Machine Translation (Chu & Wang, COLING 2018). Also, you can... Read More

In this post, I would like to introduce a survey paper titled Multi-Task Learning in Natural Language Processing: An Overview. Here, I also recommend the readers to read another survey paper on Multi-Task Learning in NLP (Multi-task learning for natural language processing in the 2020s: where are we going?) and a classical paper (Multi-Task Deep... Read More

In this post, I would like to introduce a survey paper titled A Survey of Multilingual Neural Machine Translation published on ACM Computing Surveys, which is writen by the team of NICT. Here, I also recommend the readers to read another analysis paper of MNMT (Kudugunta et al., 2019), which is also a nice paper for comparing representations ac... Read More

Survey on Low-Resource Machine Translation In this post, I would like to introduce a survey paper titled Survey of Low-Resource Machine Translation from arrive, which is writen by the team of University of Edinburgh. Current Machine Translation (MT) model are typically trained on data sets consisting of tens or even hundreds of millions of ... Read More